- Revert patch number 92. It turned out to be broken behind hope.  Personally I
   attribute this to the recent heat wave over here and apologize for the
   problems this may have caused. Turned out that my note about the change
   beeing dnagerous in the last change log was more then true...

 - Locking issues for ioctl handling.

 - Remove waiting_for_dma bit field. Use IDE_DMA bit flag instead.
   Apply this bit globally and not in the corresponding implementation
   functions.

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

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

into mandrakesoft.com:/home/jgarzik/repo/net-drivers-2.5

* fix big endia multi-cast filtering
* fix 8139too ethtool media interface

into mandrakesoft.com:/home/jgarzik/repo/net-drivers-2.5

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

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

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

So the missing EXPORT_SYMBOL() for ioremap_nocache at least found
some ISDN drivers which shouldn't have been using it in the first
place.

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

This patch adds the PCMCIA client driver for the AVM A1/Fritz!PCMCIA
ISDN cards, which has been in 2.4 for some time already.

This fixes linux/tqueue.h compile errors in 2.5.23 that were
found by Andy Pfiffer, Matthew Harrell and me.

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

Submitted by Thunder

fsync is allowed to return early if datasync is set and the I_DIRTY_DATASYNC
flags is cleared, not if either of those is true.

Submitted by Christoph Hellwig

I've got to keep my 2.4 and 2.5 patches straight!

Use bd_claim/bd_release to guarentee exclusive access to the external
log device.  I also fixed a struct block_device leak once I touched
that code (missing bdput() both in lmLogClose and the lmLogOpen
error path) and sanitized the failure path labels.

Submitted by Christoph Hellwig

into kleikamp.austin.ibm.com:/home/shaggy/bk/jfs-2.5

The pcmcia drivers haven't been updated to include the new
tqueue.h header when using tq_structs.

I noticed the breakage when compiling i82365.c for my laptop, and
took the time to modify all the drivers in that directory.

into mandrakesoft.com:/home/jgarzik/repo/net-drivers-2.5

into mandrakesoft.com:/home/jgarzik/repo/net-drivers-2.5

I suppose one can argue whether that is ugly or rather nice for
documentary purposes, but make on its own cannot figure out where a file
is supposed to live, so we have to help it.

(For the targets handled by Rules.make the situation is different, there
 Rules.make knows what's source and what's object and can add prefixes
 as necessary, so the most of the Makefiles are actually not affected by
 this kind of change)

For now, as $(obj) = $(src) = ., we only add "./", so the potential for
breakage is rather small.

This wasn't quite merged up to the current kbuild.

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

akpm pointed out that an EXPORT_SYMBOL is missing for default_wake_function.

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

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

Submitted by Manfred Spraul.

Cset exclude: torvalds@penguin.transmeta.com|ChangeSet|20020619003306|07760
Cset exclude: ak@muc.de|ChangeSet|20020618172743|19150

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

Previously each raid personality (Well, 1 and 5) started their
own thread to do resync, but md.c had a single common thread to do
reconstruct.  Apart from being untidy, this means that you cannot
have two arrays reconstructing at the same time, though you can have
to array resyncing at the same time..

This patch changes the personalities so they don't start the resync,
but just leave a flag to say that it is needed.
The common thread (mdrecoveryd) now just monitors things and starts a
separate per-array thread whenever resync or recovery (or both) is
needed.
When the recovery finishes, mdrecoveryd will be woken up to re-lock
the device and activate the spares or whatever.

raid1 needs to know when resync/recovery starts and ends so it can
allocate and release resources.
It allocated when a resync request for stripe 0 is received.
Previously it deallocated for resync in it's own thread, and
deallocated for recovery when the spare is made active or inactive
(depending on success).

As raid1 doesn't own a thread anymore this needed to change.  So to
match the "alloc on 0", the md_do_resync now calls sync_request one
last time asking to sync one block past the end.  This is a signal to
release any resources.