The talk about disconnect locking reminded me of this - the error code in
the no-driver case for the disconnect ioctl isn't a unique error code.
This changes the error code to what getdriver() uses, -ENODATA.

Here's a more complete patch for that driver binding problem.
Now (a) when scanning for a driver to handle a new device,
scanning stops after the first successful probe not the
first successful match; and (b) new drivers get every chance to
bind to additional devices, even after the first one succeeds.

Please merge.  This will help resolve some of the USB flakies
in recent kernels.  Note that (a) was noticed independently by
me and Andries Brouwer, and AFAIK this is the first patch to
also fix failure (b) which was uncovered by fixing (a).

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

Various small fixes and adds ids for new test firmware.

I think the qh needs to be removed from its qh->list, or else
uhci_remove_qh will incorrectly change the previous endpoint's qh->link,
undoing what was just done in uhci_delete_queued_urb.

Slight change - the int qh order is actually opposite of the example.

This is 95% Dan's patch, but I made some small changes. The changes I've
made relative to Dan's patch is:

Drop concept of skeleton TD. After Dan's patch, this was reduced to one
entry in the form of skel_term_td. That wasn't even a skeleton TD in the
first place. This simplifies the code a little bit.
Minor formatting tweaks.
Pass on USB bandwidth changes for now.
Pass on Interrupt auto-resubmit changes for now.
Use complete_list_lock, not complete_list as the lock.
Use frame_list_lock, not fame_list_lock.
Reorganize skeleton QH's to match the order in the schedule. This cleaned
up some debugging code and made the list more logical.
Update some obsolete documentation.
Pass on code to check for race conditions and fix up. It was racy as well.

Dan, do you compile with SMP? Those two locking typos would have
generated a warning and/or error.

I've done some light testing and haven't found any problems yet. I
haven't tested the control or interrupt queuing for lack of anything to
test with yet.

Anyway, please apply.

   Merge in Dan's patch to add interrupt and control queuing support. Summary of changes:
   Add queuing support for Interrupt and Control message in addition to Bulk.
   This resulted in some merging of code.
   Fix a queuing bug when moving a child into the parent position.
   Update documentation.
   Update debugging code.

Woops, my fault. I forgot to send you this patch which needs to be
applied before the big one.

It's from Dan as well and switches low speed control to use breadth
first traversal to make it more fair.

  Don't make low speed control use depth first. That isn't
  particularly fair. Thanks to Dan Streetman for bringing
  this up and the original patch.

This is a patchlet that updates a comment to use the correct path to the
proc debugging file.

HID-BP confuses people, we know that.
I don't want to see the HID-BP drivers removed,
but I think it would be a good idea to make them more
difficult to enable, unless someone knows what they
are looking for.  (Basically it would reduce support
incidents a lot.)

The second way puts HID-BP in its own sub-menu and
updates the Help file.  I think that this is the better
choice, so unless there is a great hue and cry,
please apply the "hidbpmenu-2542.patch" file.

This patch:
	(a) updates some comments
	(b) adds some debugging
	(c) fixes CB and CBI data stages
	(d) adds some missing {'s and }'s -- only dumb luck let the old
	version compile.

I'm applying Dan's patches by hand to double check everything and I
wanted to get a couple of the trivial ones out of the way first.

This patch fixes a bug where errors on the submit path wouldn't remove
URB's from the HC linked list.

  If we fail adding the URB to the schedule, we need to make
  sure that we remove it from the urb_list. Thanks to
  Dan Streetman for finding and fixing this bug.

This patch fixes a typo in the names used to remove the proc directory
that uhci uses.

The longer term fix of splitting out (or dropping outright) the
debugging code will come soon after the important bugs and features are
merged.

> I've seen several cases where 'usbtest' would look at the device
> and say "no thanks", but then the "right" driver wouldn't ever
> bind to the device.

This patch seems to fix it.  I think Linus' tree needs it,
or something like it.

(Since the probe routine returns zero on success, there
was confusion inside the driver model code where it assumed
that zero meant failure.)

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

This patch from Christoph Hellwig removes the kiobuf/kiovec
infrastructure.

This affects three subsystems:

video-buf.c:

   This patch includes an earlier diff from Gerd which converts
    video-buf.c to use get_user_pages() directly.

   Gerd has acked this patch.

LVM1:

   Is now even more broken.

drivers/mtd/devices/blkmtd.c:

   blkmtd is broken by this change.  I contacted Simon Evans, who
   said "I had done a rewrite of blkmtd anyway and just need to convert
   it to BIO.  Feel free to break it in the 2.5 tree, it will force me
   to finish my code."

Neither EVMS nor LVM2 use kiobufs.  The only remaining breakage
of which I am aware is a proprietary MPEG2 streaming module.  It
could use get_user_pages().

From Anton Blanchard.  This fixes a couple of Linux Test Project
failures.

- Returns EBUSY if the caller is trying to invalidate memory which is
  covered by a locked vma.

  The open group say:

  [EBUSY]
          Some or all of the addresses in the range starting
          at addr and continuing for len bytes are locked,
          and MS_INVALIDATE is specified.

- Returns EINVAL if the caller specified both MS_SYNC and MS_ASYNC

  [EINVAL]
          The value of flags is invalid.

  and:

          "Either MS_ASYNC or MS_SYNC is specified, but not both."

CONFIG_LBD was initially hardwired to "on" for testing.  That seems to
have gone OK, so now make it configurable.

From Bill Irwin: ensure that the data structures which hold the node's
zone structures is all zeroed before we start using it.

If you're peforming 512-byte sized IOs.  With, say,

	dd of=/dev/raw/raw1 bs=512

then the `pgpgin' and `pgpgout' accounting just sits on zero.

This is because it counts in kbytes, and 512/1024 is zero.

So change it to count sectors, and divide that by two when we report it
to userspace.

From Ed Tomlinson, then mauled by yours truly.

The current shrinking of the dentry, inode and dquot caches seems to
work OK, but it is slightly CPU-inefficient: we call the shrinking
functions many times, for tiny numbers of objects.

So here, we just batch that up - shrinking happens at the same rate but
we perform it in larger units of work.

To do this, we need a way of knowing how many objects are currently in
use by individual caches.  slab does not actually track this
information, but the existing shrinkable caches do have this on hand.
So rather than adding the counters to slab, we require that the
shrinker callback functions keep their own count - we query that via
the callback.

We add a simple registration API which is exported to modules.  A
subsystem may register its own callback function via set_shrinker().

set_shrinker() simply takes a function pointer.  The function is called
with

	int (*shrinker)(int nr_to_shrink, unsigned int gfp_mask);

The shrinker callback must scan `nr_to_scan' objects and free all
freeable scanned objects.  Note: it doesn't have to *free* `nr_to_scan'
objects.  It need only scan that many.  Which is a fairly pedantic
detail, really.

The shrinker callback must return the number of objects which are in
its cache at the end of the scanning attempt.  It will be called with
nr_to_scan == 0 when we're just querying the cache size.

The set_shrinker() registration API is passed a hint as to how many
disk seeks a single cache object is worth.  Everything uses "2" at
present.

I saw no need to add the traditional `here is my void *data' to the
registration/callback.  Because there is a one-to-one relationship
between caches and their shrinkers.


Various cleanups became possible:

- shrink_icache_memory() is no longer exported to modules.

- shrink_icache_memory() is now static to fs/inode.c

- prune_icache() is now static to fs/inode.c, and made inline (single caller)

- shrink_dcache_memory() is made static to fs/dcache.c

- prune_dcache() is no longer exported to modules

- prune_dcache() is made static to fs/dcache.c

- shrink_dqcache_memory() is made static to fs/dquot.c

- All the quota init code has been moved from fs/dcache.c into fs/dquot.c

- All modifications to inodes_stat.nr_inodes are now inside
  inode_lock - the dispose_list one was racy.

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

Cleaning out my queue of most minor patches:

  - Provides some kerneldoc for 'struct usb_interface' now that
    the API is highlighting it.

  - Fixes usb_set_interface() so it doesn't affect other interfaces.

    This provides the right place for an eventual HCD call to clean
    out now-invalid records of endpoint state, and also gets rid of
    a potential SMP issue where drivers on different interfaces
    calling concurrently could clobber each other.  (Per-interface
    data doesn't need locking except against config changes.)

  - It's OK to pass URB_NO_INTERRUPT hints if you're queueing a
    bunch of interrupt transfers.

The set_interface call should eventually take the interface as a
parameter, it's one of the few left using the "device plus magic
number" identifier.  I have a partial patch for that, but it doesn't
handle the (newish) ALSA usb audio driver or a few other callers.

Thanks to David Brownell for pointing this out to me.

Here is patch which adds parallel link cable support for Texas Instruments
graphing calculators.

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

Keep the partition information around for the lifetime of the module.

This updates AFS mtd partition parsing to the current CVS version:
- Don't recognise the AFS SIB as a partition
- Ensure initialisation of afs mtdpart structures.

We removed asm/mach-types.h from asm/hardware.h.  This means we must
now include asm/mach-types.h where its used.

We also fix h3600 pcmcia initialisation/cleanup.

Traditionally, the Assabet reverses its mapping of UART1 and UART3
when the Neponset board is connected.  This can be (a) confusing
and (b) annoying when the boot loader uses UART1.  We therefore
have a fixed mapping between the ttySA names and the physical
UARTs on this platform.

The patches convert drivers/pcmcia to use C99 named initializers,
and all the patches are against 2.5.42. There are 25 patches in
total, and the "cat"ing them together they're more that 20K, so
I'm sending the patches as a compressed attachment. The patches
were CC'd to Linus in the first mail that bounced.

Here's a small set of patches that switch the code to use C99
desiginated initializers. Patches are against 2.5.42.

Prevent the assembler putting constant pools in the middle of code.
Clean up shark ISA PIC handling.

This cset removes the global irq handling in the AcornSCSI driver,
and makes the target type for bitops an unsigned long array rather
than an unsigned char array.