This fixes the scheduler's migration code to not disable preemption.  It
also fixes the bug that was hidden by the broken disable-preempt change:
the migration thread did not kick idle CPUs if a task is migrated to
them, which causes a hung boot when ksoftirqds are started.  It was pure
luck it worked until now, it was broken pretty much from day 1 on.

Quick fix for CONFIG_NLS, add a missing space.

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

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

This is a minimal patch to allow me to load/use the atp module so I can do
further testing work on it.

Here's an additional patch that contains the cleanups I did to John's
timer patches.  It does the following:

	- uses C99 initializers
	- makes the timer list static
	- adds better documentation to the timer function structure
	- makes the timer init function return 0 on success
	- NULL terminates the list of timers to make further patches
	  easier.

The 2.5.x RPC code is currently broken in that it demands that all
tasks that call xprt_create_proto() in order to open a TCP socket must
have CAP_NET_BIND_SERVICE capabilities, and must bind to a privileged
port.

This breaks the NLM locking code and its use of the call_bind() RPC
portmapper lookup feature.

This patch allows the built-in portmapper client to use unbound TCP
sockets if the user does not have the necessary capabilities.

This is a nontrivial change to the NFS client.

In this patch, we finish modifying the async READ path so that it is
version-agnostic.  We define a new nfs_rpc_op ->setup_read(), and move
the v2- and v3-specific code in nfs_read_rpcsetup() there.  We also
have to change nfs_readpage() result so that the 'count' of bytes
read is a parameter.  The extra parameter means that it can no longer
be ->tk_exit().  Instead, it is called from a version-specific ->tk_exit()
routine which is set in ->read_setup().

The upshot of all this is that the version-specific part of the
async READ path has been encapsulated in a new nfs_rpc_op
->read_setup(), and NFSv4 can share the logic for asynchronous
READ's with NFSv2 and v3.

This is a nontrivial change to the NFS client.

Synchronous READ operations are currently done via the ->read() nfs_rpc_op.
Therefore, the synchronous READ path can easily be adapted for NFSv4.  On
the other hand, the asynchronous READ path contains several NFSv3-specific
features, which make it difficult to adapt for NFSv4.

In this patch and the next, we modify the async READ path to be
version-agnostic.  This patch just changes the 'struct nfs_read_data'
so that the v2- and v3-specific parts are moved into a private area,
with room for a v4-specific part in parallel.  None of the logic is
changed.

If the NFS_ATTR_FATTR_V4 flag is set, use the NFSv3 convention for
the 'space_used' part of the fattr.

This is a nontrivial change to the NFS client.

NFSv4 defines a new file attribute, change_attr.  This is a per-file
opaque quantity returned by the server, whose value is required to
change whenever the file is modified.  If it exists, we want to use
it for all cache consistency checks in nfs_refresh_inode().  Some
operations also return a "pre-operation" value of the change_attr;
we want to take this into account too.

First, define flags
  NFS_ATTR_FATTR_V4 - indicates that the 'struct nfs_fattr' is an
                      NFSv4 fattr, so the change_attr field is valid
  NFS_ATTR_PRE_CHANGE - indicates that the server returned a pre-operation
                      change_attr, so the pre_change_attr field is valid

Second, change nfs_refresh_inode() so that the caches are invalidated
if there is a change_attr mismatch.  Exception: If the pre_change_attr
tells us that the mismatch was caused by our operation, then do not
invalidate the caches.

This patch should leave the logic in nfs_refresh_inode() unchanged
if neither of the new flags are set.

In NFSv4, an fsid is a 64-bit major number together with a 64-bit
minor number.  In previous versions, an fsid is a single number.
This patch changes 'struct nfs_fattr' accordingly.

In NFSv4, there is no hard limit on the length of symlink text.
This patch changes the -ENAMETOOLONG test in nfs_symlink() accordingly.

This patch changes the interface of the ->setattr() nfs_rpc_op
so that its first argument is a dentry instead of an inode.

[Explanation: The dentry is required because in NFSv4, we may
 need to OPEN the file before doing the SETATTR.  (This is
 required if the file size is changed as part of the setattr.)
 Opening the file requires making use of the containing
 directory's inode.]

This patch changes the interface of the ->readdir() nfs_rpc_op
so that its first argument is a dentry instead of an inode.

[Explanation: The dentry is required because in NFSv4, we need
 to make use of the _parent_ directory's inode.  This is because
 NFSv4 servers no longer return an entry for ".." in the READDIR
 response, so the client kernel needs to fake this entry, inode
 number and all.]

In a number of places in the NFS client, I had to change

  #ifdef CONFIG_NFS_V3
     /* ... */
  #endif

to

  #if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4)
     /* ... */
  #endif

Instantiate a new file, include/linux/nfs4.h, which contains
constants and typedef's for the NFSv4 protocol (by analogy with
include/linux/nfs2.h and include/linux/nfs3.h).

Also #include this file in a few places where it will be needed
later.

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

The patch below removes some dead code and nonsense code.
The part that changes behaviour is

 -       if (sbi->s_cruft == 'n' &&
 -           (volume_seq_no != 0) && (volume_seq_no != 1)) {
 -               printk(KERN_WARNING "Warning: defective CD-ROM "
 -                      "(volume sequence number %d). "
 -                      "Enabling \"cruft\" mount option.\n", volume_seq_no);
 -               sbi->s_cruft = 'y';
 -       }

that has already bitten lots of people.

Nothing is wrong with a volume sequence number different from 0 or 1.
(Cf. Ecma-119.pdf, Sections 4.17, 4.18, 6.6.)

This patch takes out the unused KERN_PPC_ZEROPAGED sysctl, and
restricts the KERN_PPC_POWERSAVE_NAP and KERN_PPC_L2CR sysctls to be
present only on those PPC processors where they are useful.  This
patch only affects PPC.

This patch adds the PCI device ID for the Motorola MPC107 host bridge.
The entry is already in the list at pciids.sf.net but isn't in the
kernel pci_ids.h file yet.  Please apply this to your tree.

This patch allows to change uid, gid and mode of files and directories
located in procfs.

Without this patch you can change uid, gid and mode as long as the
file is open. As soon as you close the file, it reverts back to its
default, which is root:root and readonly usually.

Reference to .ver file incorrect after recent makefile changes.
Grepped the kernel tree, and this is the only Makefile that
uses $(MODVERDIR).

CD-ROM puts struct cdrom_changer_info on the stack in a few places, this
is a bad idea since it's big (a bit over 1kb).  This makes us allocate
it instead.

Noticed by Anton.

into the proper subdirectory (kernel) where it is used.

Drop unused variables.

        In order to demonstrate how new time-sources are added to my
timer-changes patch. Here is my current version of my cyclone-timer
patch for 2.5.41. This uses the infrastructure set up in the
timer-changes_A4 patch set to add the cyclone counter (found on IBM
Summit Based hardware) as a time-source.

        The current code is not enabled as it also depends on James
Cleverdon's 2.5 summit patch, however it illustrates how cleanly new
time-sources can be added.

        This is the final part 3 of 3 of my timer-change patch. Part 3
integrates the moved code (from part 2) into the new infrastructure
(from part 1).

        This is part 2 of 3 of my timer-change patch. Part 2 is just a
bulk move of code out of time.c and into timer_pit.c and timer_tsc.c. No
code is changed, only moved.

        Please note, this code will not compile without the final third
part of this patch collection. This was done for readability alone.

        The i386 time.c code is turning into a mess. We've got multiple
functions that do the same thing, only with different hardware, all
surrounded #ifdefs and even more difficult to follow #ifndefs. George
Anzinger is introducing a new ACPIpm time source, I'm going to attempt
to add the cyclone counter as a time source, and in the future there
will be HPET to deal with. These will not go in cleanly together as
things are now.

        Inspired by suggestions from Alan, this collection of patches
tries to clean up time.c by breaking out the PIT and TSC specific parts
into their own files. Additionally the patch creates an abstract
interface to use these existing time soruces, as well as make it easier
to add future time sources.

        It introduces "struct timer_ops" which gives the time code a
clear interface to use these different time sources. It also allows for
clearer conditional compilation of these various time sources.

        This first patch (part 1 of 3) provides the infrastructure
needed via the timer_ops structure, as well as the select_timer()
function for choosing the best available timer.

Should resolve an ongoing fiasco concerning what we should return to
userspace if they do a readv or writev of zero segments.

SuS is ambiguous, but implies EINVAL.  We're currently returning
EINVAL, but 2.4 returns zero.

I think zero makes more sense, and it is what 2.4 does.

The changed sched_yield() semantics have made ext3's transaction
batching terribly slow.

Apparently a schedule() fixes that, although it probably breaks
transaction batching.

This patch largely fixes my complaints about the new scheduler being
extremely sluggish to interactive applications.  Evidently those
applications were calling fsync() and were spending extremely long
periods in sched_yield().

From Hugh Dickins.

radix_tree_reserve() exists solely for the tmpfs move_to_swap_cache()
and move_from_swap_cache() functions, and yet they don't need it: there
is no problem in the one page being simultaneously listed in two radix
trees (while both locks are held).  Use radix_tree_insert(), and remove
radix_tree_reserve(); also removed a few blank lines.

Fix the raw driver by tricking it into performing O_DIRECT IO against
the bound blockdev.

- rewrite the i_mapping for /dev/raw/raw0 to point at the same thing
  as bdev->bd_inode->i_mapping.  We've performed a bdget() against the
  blockdev, which should pin it for the correct lifetime.

- set the O_DIRECT bit on the caller's file->flags.

The atomicicty fix for move_one_page() was not quite right.

We only do the page_table_present() test if CONFIG_HIGHPTE=y.  Which is
fine, but even with CONFIG_HIGHPTE=n, the pte mapping functions still
do an inc_preempt_count() due to their unconditional kmap_atomic().  So
we get a might_sleep() warning.

The warning is actually bogus, because those pte's are always in
direct-mapped memory.

So hm.  Three fixes suggest themselves:

1: Run the page_table_present() test if CONFIG_HIGHMEM.

   Rejected: penalises non-pte_highmem setups

2: Make kmap_atomic() not do inc_preempt_count() is the page was
   direct mapped.

   Rejected: I don't think we want kmap_atomic side effects to be
   varying according to the page which was passed.

3: Change the pte mapping functions so they don't run kmap_atomic at
   all if CONFIG_HIGHPTE=n

   This is what I did.  And guess what?  For CONFIG_HIGHMEM=y,
   CONFIG_HIGHPTE=n this patch shrinks the kernel by 5 kbytes.  Because
   kmap_atomic is inlined.

   The lesson: we do way too much damn inlining.

I have invented a new software development methodology! You send an
email to Hugh saying "I don't have the foggiest idea why this guy's
kernel is oopsing" and next morning, you get a patch! I shall patent
this.

Since 2.5.3, move_vma() has been passing a freed vma into
move_page_tables().  Fix it to move back to the previous vma in the
list if we're about to delete this one.

Thanks to Morten Helgesen for patient reporting, diagnosis and testing.

This is my latest timer patchset, it makes del_timer_sync() a bit more
robust wrt.  code that re-adds timers from the timer handler.

Other changes in the patch:

 - clean up cascading a bit.

 - do not save flags in __run_timer_list - we enter from an irqs-enabled
   tasklet.