From: Trond Myklebust <trond.myklebust@fys.uio.no>


knfsd needs to disable soft interrupts when calling
csum_partial_copy_to_xdr().

At the moment there's a nasty conflict between the RPC server and
client. The problem arises when you get to xdr_partial_copy_from_skb()
(and the kmap_atomic()); the RPC client can end up calling the same
function from a ->data_ready() soft interrupt, and corrupt any data
the knfsd process may have copied.

Superblock format '1' resolves a number of issues with
superblock format '0'.
It is more dense and can support many more sub-devices.
It does not contains un-needed redundancy.
It adds a few new useful fields

The code to understand a specific superblock format is
already highly localised in md.  This patch defines a
user-space interface for selecting which superblock format
to use, and obeys that selection.

Md currently has a concept of 3 version numbers:
 A major version number
 A minor version number
 A patch version number

There historically seems to be some confusion about whether
these refer to a version of the superblock layout,
or a version of the software.

We will now define that:
   the "major_version" defines the superblock handler.
      '0' is the current superblock format.  All new formats
          will need new numbers.
   the "minor_version" can specify minor variations in the
      superblock, such as different location on the device
   the "patch_version" will be used to indicate new extenstions
      to the software.. patch_version=1 will mean multiple superblock
      support.

A superblock version number is selected by specifing major_version
in SET_ARRAY_INFO ioctl.

This patch:
  Updates Documentation/md.txt with details of new interface.
  Generalises desc_nr handling and makes sure that an array never
  has two devices with the same desc_nr.
  makes sure mddev->major_version is always valid and is 0 by default.
  uses mddev->major_version to select superblock handlers.
  Modifies set_array_info to just record version number if raid_disks==0
  Makes sure max_disks is always set correctly.
  Determines device size when reading superblock, or a hot-add/add-new.

Normally the data stored on a component of a RAID array is stored
from the start of the device.  This patch allows a per-device
data_offset so the data can start elsewhere. This will allow
RAID arrays where the metadata is at the head of the device
rather than the tail.

Md devices (raid1/raid5) can resync or recover.
There are similar but importantly different.
resync happens after an unclean shutdown
recovery happens when a failed drive is being replaced by a hot spare.

The sync-checkpoint code confused the two somewhat and this causes
problems.

This patch makes sure "recovery_cp" only relates to resync,
not recovery.
It also fixes a small problem with recording spares in
the superblock.

From: Angus Sawyer <angus.sawyer@dsl.pipex.com>

If there are no writes for 20 milliseconds, write out superblock
to mark array as clean.  Write out superblock with
dirty flag before allowing any further write to succeed.

If an md thread gets signaled with SIGKILL, reduce the
delay to 0.


Also tidy up some printk's and make sure writing the
superblock isn't noisy.

The md_recoveryd thread is responsible for initiating and cleaning
up resync threads.
This job can be equally well done by the per-array threads
for those arrays which might need it.

So the mdrecoveryd thread is gone and the core code that
it ran is now run by raid5d, raid1d or multipathd.

We add an MD_RECOVERY_NEEDED flag so those daemon don't have
to bother trying to lock the md array unless it is likely
that something needs to be done.

Also modify the names of all threads to have the number of
md device.

It is needed for kernel_fpu_*

Md uses ->recovery_running and ->recovery_err to keep track of the
status or recovery.  This is rather ad hoc and race prone.

This patch changes it to ->recovery which has bit flags for various
states.

It is (now) too trivial to even have an inline.

From: Art Haas <ahaas@airmail.net>

This patch converts the file to use C99 initializers.

From: Angus Sawyer <angus.sawyer@dsl.pipex.com>

Mainly straightforward convert of sprintf -> seq_printf.  seq_start and
seq_next modelled on /proc/partitions.  locking/ref counting as for
ITERATE_MDDEV.

pos == 0 -> header
pos == n -> nth mddev
pos == 0x10000 -> tail

Whenever a ITERATE_MDDEV loop is exitted abnormally
we need to mddev_put the current mddev.  There was
one point in md_do_sync where we didn't so use counts
became wrong.

set. A few versions of gcc-2.96 generate seriously incorrect code.

This patch creates a new include directory include/media, populates
it with a few files header files and fixups the affected drivers to
compile with the new directory layout.

The directory is intented to be used for (kernel-internal) header files
of the media drivers (which are sitting below drivers/media).  For now
the video-buf.h (mm helper), tuner.h (tv/radio tuner) and audiochip.h
(tv sound decoder drivers) header files are moved.  Some more header
files from the dvb folks will likely follow.

This patch deletes the MIN/MAX macros from audiochip.h
and fixes all users of these macros to use the kernels
min/max macros instead.

Patch from Neil Brown <neilb@cse.unsw.edu.au>

Allow auto-unplugging to work for devices that do it themselves.

Auto-unplugging - and blk_unplug_work in particular - assumes that the device
uses "generic_unplug_device" for unplugging, but some devices don't.  md
crashes.

So blk_unplug_work should use ->unplug_fn

This patch fixes spelling errors in the Documentation directory and was
extensively reviewed by the following people: Dan Kegel, David
Woodhouse, Joe Perches, Jared Daniel J.  Smith, Michael Hayes, and
Steven Cole

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

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

into kernel.bkbits.net:/home/davem/net-2.5

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

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

From Ben Collins

Used as manual trigger for device/driver binding, so drivers can claim 
devices that are present but failed to be bound. In case of ieee1394, this
can happen if two computers are attached to a device - the second one won't
be able to claim it until the first detaches. 

Changed device_attach() to return 1 when a driver was bound to a device. 
Makes it so bus_rescan_devices() can return the number of devices attached.
Tested this with ieee1394 over a bus reset and it worked as expected.

- Don't d_delete() files when removing a directory - they must still be 
  hashed when dput() reaps them, and it will __d_drop() them anyway.

- Don't d_invalidate() the directory on removal. 

- do simple_rmdir() after d_delete() on directory. 

- Remove extra dput(), which was causing refcount to go negative, causing 
  an oops when someone tried to create the directory again.

- Make sure we don't d_delete() or do extra dput() on file when updating, 
  either.

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

The old adap_lock and driver_lock could deadlock and there was a bunch
of stuff that was horribly racy.  The new locking scheme is from
lm_sensors CVS but I had to fix quite a few bugs to make this work.

This one is from lm_sensors CVS and mostly cosmetic changes.  I it up a
bit to compile properly under latest 2.5.

another bunch of IDs appeared in lm_sensors CVS lately..

This somehow got lost in the uClinux merge.

It's the _device_ filesystem, so regular files are grossly misplaced
here.  Fortauntely only one driver actually tries to use it (microcode)
and it's works fine with  a regular miscdevice as well.

Sorry, the devs_only() patch accidentally included the devfs-portion
of the DEVFS_FL_AUTO_DEVNUM patch.  This patch contains the remaining
parts.

The new slab poisoning code broke kmem_cache_size(), it now returns
a too large size as the poisoning area after the object is includes.
XFS's kmem_zone_zalloc thus overwrites exactly that area and triggers
the new checks everytime such an object is freed again.

I don't recommend using XFS on BK-current without this patch applied :)