This is the only user I'll feed you this time. As with 2.4 I want it to
run for a bit on read only media first 8)

We need to copy the new io op. Actually we need to shoot the entire mess
in this function and make the drivers always set it but not today.

(is_flashcard is unneeded outside ide_probe as we have a drive->flash check)

Some drivers need this. Its in the core as the core eventually needs to
be doing the tracking here

The iops struct may be a good idea in the longer run but right now its
unused and its mess that can be restored neatly later on.

(second hunk just makes 2.4/2.5 header match format)

Now handled at runtime

We cannot do an identify on a drive with no driver loaded. The kernel
has a ton of half backed "if no driver" cases but they dont cover all
cases and its a mess.

For now we rely on the probe time identify unless a driver is loaded. A
proper fix (an 'ide-unassigned' driver) will follow later

Fix crash with slave and no master
Fix crash with hdb=noprobe hdb=cdrom
Fix crash with pre ATA devices refusing IDENTIFY
Fix flash slave making master disappear
Add interfaces that the PPC uses to do disk spin up when the BIOS has not
Add framework to allow hdparm -d1 on a box built with IDE_DMA_ONLYDISK
Fix identify framework so we can fix the proc identify crash too

Also adds the new OUTBSYNC iop

Remove ide_ioreg_t
Add locking on the ide setting lists

This abstracts out the mmio copies as PPC at least has better ways to
this and there are other issues on other platforms. It keeps DaveM happy
too 8)

Ok this first piece eliminates some of the use of ide_ioreg_t, which
actually only works as a ulong anyway. Its a dysfunctional abstraction.

We also need the ide pci stuff if we have IDE PCI not if we have PCI but
only legacy mode IDE support

There is no reason to have both BSD_PARTITION and FREEBSD_PARTITION
denoting the same partition type.

When a raid1 or raid5 array is in 'safe-mode', then the array
is marked clean whenever there are no outstanding write requests,
and is marked dirty again before allowing any write request to
proceed.

This means than an unclean shutdown while no write activity is happening
will NOT cause a resync to be required.  However it does mean extra
updates to the superblock.

Currently safe-mode is turned on by sending SIGKILL to the raid thread
as would happen at a normal shutdown.  This should mean that the
reboot notifier is no longer needed.

After looking more at performance issues I may make safemode be on
all the time.  I will almost certainly make it on when RAID5 is degraded
as an unclean shutdown of a degraded RAID5 means data loss.

This code was provided by  Angus Sawyer <angus.sawyer@dsl.pipex.com>

This allows the thread to easily identified and signalled.
The point of signalling will appear in the next patch.

From Art Haas <ahaas@airmail.net>:

This converts md.c to use C99 initializers to improve readability and
remove warnings if '-W' is used.

by the nature of multipath, already be in-sync, so we should set the
in-sync flag.

From Andrew Morton <akpm@digeo.com>:

gcc-2.95.3 is getting an internal compiler error with CONFIG_LBD=y.  Reorganising
the code a bit made it go away.

This call the super_90_load is meant to make sure that the new
superblock is consistant with a pre-exisitnig one (on rdev0)...  but
rdev0 was not passed :-(

From Art Haas <ahaas@airmail.net>:

This converts the file to use C99 named initializers.  These change make
the file compile with fewer warnings if '-W' is added to the compile
flags, and may enhance code readability.  Let me know if you think this
should be sent to Linus.

From Trond:

  When the nanosecond resolution on [acm]time was introduced to 2.5.x,
the knfsd GETATTR responses were converted to make use of the
nanosecond field, but the pre/post WCC attributes were not. This will
lead to a lot of unnecessary cache invalidations on the clients.

The following trivial patch should fix up knfsd so that it stores and
encodes the full 'struct timespec' in both pre and post attribute
fields.

Problem was that the cache was being updated inplace, rather
than swapping in a new entry, so old filesystem pointers
were overwritten without being released.

As nfsv4 does some reply encoding while processing
requests, the code which tries to put the overall status
at the head of the reply fails.

With this patch, we take the address for the overall
status very early so NFSv4's encoding wont confuse us.

NFS replies can have a 'tail' after a 'data' component
(for read, readlink, readdir).

For nfsv2 and v3 this tail is in the same page as the head.
For nfsv4, this tail is in a separate page.

We need to keep track of which page so that it right one
gets sent.  This patch does that.

From Ted Phelps <phelps@dstc.edu.au>

The NFSv3 readdirplus path doesn't check to see if maxcount is less
than the size of a page before it fills it up, possibly overwriting
random bits of memory.  At least, it makes my Solaris NFSv3 client
work.

The attached patch, against 2.5.58, adds this check in a way which is
consistent with the way readdir does under both NFSv2 andNFSv3.