Bug 516.

Use schedule_delayed_work instead of a timer should fix this. Thanks
to Andrew Morton and Russell King.

(Added flush_scheduled_work() to ensure our delayed work completes
before we free the pcmcia_bus_socket structure. --rmk)

The socket drivers already offer suspend and resume
capability. Integrate this with the driver model, based on a
suggestion by Russell King.

Also, remove two never-used functions from the socket drivers (to_ns).

 drivers/pcmcia/cs.c             |   70 ++++++++++++++++++++--------------------
 drivers/pcmcia/cs_internal.h    |    1
 drivers/pcmcia/hd64465_ss.c     |    2 +
 drivers/pcmcia/i82092.c         |   17 ++++++---
 drivers/pcmcia/i82365.c         |    2 +
 drivers/pcmcia/pci_socket.c     |   15 +-------
 drivers/pcmcia/sa1100_generic.c |    2 +
 drivers/pcmcia/sa1111_generic.c |   14 +-------
 drivers/pcmcia/tcic.c           |    7 +---
 include/pcmcia/ss.h             |    5 ++
 10 files changed, 64 insertions(+), 71 deletions(-)

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

Don't allow "bind_request" to be called before "register_pccard_driver".

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

This fixes ipmi compilation (with some documentation updates and another
minor fix, too).

It also fixes a problem with the state machine getting stuck in a
certain error condition.

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

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

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

we're writing over TCP and have done a partial send.

  The following patch by Joe Korty removes an over-zealous check in the NFS
  read code that causes pages to be incorrectly marked with PG_error and
  hence causes an EIO to be returned to userland.

  The test is incorrect as it ignores the fact that we may be caching a
  write that will extend the file on the server (and hence will create a
  hole in the region concerned.).

a null 'p_proc' procedure pointer.

and drop the 'flags' argument (it was always set too LOOKUP_PARENT).

If the user supplies and empty path, return -ENOENT.

Doing this ensures that we keep probing the connection to the server
rather than just waiting for the entire congestion window to time out.
The latter can be very expensive due to the exponential backoff rule.

Thanks to  Angus Sawyer <angus.sawyer@dsl.pipex.com> and
   Daniel McNeil <daniel@osdl.org>

Also reformat some printks so lines don't exceed 80chars.

Thanks to  Angus Sawyer <angus.sawyer@dsl.pipex.com>

From: Daniel McNeil <daniel@osdl.org>

This fixes an oops caused by incorrect usage of sector_div()
in which_dev() in md/linear.c.  It was dereferencing an non-existent
hash table entry.

We currently hold a read_lock of dparent_lock
while calling exp_get_by_name on several ancestors
of a given dentry.  However exp_get_by_name can
malloc(GFP_KERNEL), so that isn't a good idea.

Now we only claim the lock while actually
stepping up the parent chain.

This addresses bug 29 @ bugme.osdl.org

encode_read: change 'len' variable from unsigned long to a long for
'(len > 0)' comparison.  don't set up a tail iovec for zero length
reads.

NFSv4 operations after WRITE are decoded into wr_vec[] pages, thus the
argp->pagelen can be non-zero at the end of decoding the WRITE
operation.

This patch correctly sets argp->pagelen, and correctly advances argp->p
after the WRITE operation

Fix some misleading comments, and correct the test for requesting more
than one page.

From Steven Cole

From Osamu Tomita

(Stanford Checker, Chris Wright)

From Christoph Hellwig I think