Patch from Stephen Cameron <steve.cameron@hp.com>

Make cciss driver retry 3rd party bus reset aborted commands up to 3 times.
(ported to 2.5 by me, original patch by Charles White)

This is needed for a multi port storage box that can have multiple hosts
connected to it, or be used in a multipath configuration.  In certain
configurations SCSI bus resets initiated by one host may affect another host.

Patch from Stephen Cameron <steve.cameron@hp.com>


* Make cciss driver wait longer for board to enter simple mode to
  handle an unlikely corner case.  (Hot replace of failed 144GB
  drive in RAID 5 set at just the wrong time prior to driver loading
  can make the board take a longer than usual time to go into "simple
  mode".  Without the patch, the driver gives up too early, and consequently
  doesn't work.  (A reboot will generally "fix" it.)  This patch avoids
  the problem.
* Fix a couple ioctls to return EAGAIN instead of inappropriate EFAULT.

Patch from Rohit Seth

ia64 reserves virtual address region 4 (any address which starts with 0x4)
for huge pages.

Apparently, for hardware reasons, we do not wish to allow mappings with
other page sizes to appear in that region.

This patch prevents the user from being able to place regular MAP_FIXED
mappings into region 4 on ia64.  It is a no-op for ia32.

Patch from Rohit Seth.  Updates the hugetlb page documentation.

Patch from Rohit Seth.  Remove a dead declaration.

Patch from Christoph Hellwig <hch@sgi.com>

There's a bunch of minor fixes needed to disable the swap code for systems
with mmu.

Patch from "Vitezslav Samel" <samel@mail.cz>

Obvious fix of printk level (wrong since 2.5.62).

Forward port of a change which Ted made to 2.4's ext2.

HTREE backwards compatibility patch.

"I thought (and assumed) this patch had been applied to both the ext2 and
 ext3 filesystems in the 2.4 kernel.  It turns out it had only made it into
 the ext3 filesystem code.  This means that if an HTREE-enabled filesystem is
 mounted using ext2, it will corrupt the filesystem as far as e2fsck and an
 ext3 htree-enabled kernel is concerned.  (The corruption won't cause any
 data loss, but it will cause e2fsck and an ext3-htree kernel to omit a lot
 of warning messages.)"

Patch from Matthew Dobson <colpatch@us.ibm.com>

Consolidate alloc_pages() and alloc_pages_node().  There is no change in
sizeof(vmlinux).

Patch from: john stultz <johnstul@us.ibm.com>

The kernel will currently panic if it was built with CONFIG_X86_TSC and
someone boots it with `notsc'.

Change it to just print a warning.

Patch from Thomas Schlichter <schlicht@uni-mannheim.de>

Based on a patch from Dave Jones.

It converts a large number of instances of:

	smp_call_function(foo);
	foo();

into

	on_each_cpu(foo);

and in doing so fixes up the preempt-unsafeness of the first version.

Patch from Maneesh Soni <maneesh@in.ibm.com>

It's supposed to return an ERR_PTR value on error, not NULL.

The loop driver takes a copy of the data which it is writing.  When this
happens on the try_to_free_pages() path, loop can easily consume ALL memory
and bio_copy() will fail to allocate a page.

Loop forgets to check the bio_copy() return value and oopses.

Fix this by dropping PF_MEMALLOC and throttling to the block writeout speed.

The patch exports blk_congestion_wait() to modules for this.  This is a
needed export: several filesystems have a "try to allocate and yield if it
failed" loop and blk_congestion_wait() is a more appropriate way of
implementing the sleep in this situation.

Patch from Nick Piggin <piggin@cyberone.com.au>

Closes a window in which we can get an invalid (already dispatched) request
in the next_drq cache.

This fixed an oops which was reported by Janet Morgan.

Patch from Anton Blanchard <anton@samba.org>

e100 is performing a header checksum calculation which is

a) duplication of the core kernel's calculation of the same and

b) incorrect on big-endian machines.

This fix has been tested on ia32 and pppc64 and acked by davem, jgarzik
and Scott Feldman

s/seperate/separate/

Fix socket memory mapping bug - we need to take account of the
requested card offset when returning the address for ioremap. 

Request PCMCIA socket resources without marking them busy - this
allows future modifications to pcmcia to allow it to claim the
CIS resources for statically mapped sockets.

Sanitise our pcmcia interrupt handling.  Make sure we only use
set_irq_type after the interrupt has been registered.

Register the socket handling interrupts using the function name
rather than some pointer passed by the generic code.

Let the generic sa11x0 code handle the cards interrupt masking.

Remove get_irq_info method, passing the information via the init
method instead.  The IRQ allocated to a socket never changes on
these platforms, so the extra code was just bloat.

Convert all SA11x0 drivers to register with the Linux driver model,
so they work with the pcmcia layer in 2.5.63.

Make low level socket_status() method take the socket index as an
argument in the same way that our other methods do.  socket_status()
now only returns the status of the requested socket, not all sockets.

Make low-level configure_socket() method take the socket number as
a parameter rather than embedded in a structure.

Make SA11x0 generic pcmcia implementation fit better with Dominik's
work - pass around our socket structure rather than the socket number
itself.  Reference socket operations using the socket structure
rather than a static variable.

You wouldn't expect that an innocent two-line patch would practically
destroy the entire functionality of the USB subsystem.  But that's what my
last patch did.  I didn't realize until I saw it in Greg's list of patches
forwarded to Linus; that patch includes a mistakenly unreverted prior
change.  This explains the problems I was having yesterday, and probably
also the problems a great many people are having today.

Anyway, here's the correction -- it puts things back to what they should
have been in the first place.  Greg, please apply this and send it to
Linus for his tree as soon as reasonably possible.

This is a temporary patch which should get cardbus working again.
In order to use pci_remove_behind_bridge(), we need to separately
allocate the pci_dev structures.

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

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

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

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

1) Define comma so dependency files does not include a '_' in the filename
2) Use correct path for mkspec and mkversion
3) Removed checkhelp - corresponding perl script is no longer present
4) Spelling correction (reported on lkml)

For some people (though not me), the '+' indicating that a command will
invoke a sub-make didn't propagated properly, and caused a warning.
Putting the command all into one line should fix that.

Plus some cosmetics and clean up the per_cpu check.

Alas no rave reviews on lkml, but the private feedback I have received so
far on the recent changes has been excellent. The Rhine-II is now finally
usable with via-rhine. Time to call it 1.17.

This trivial patch fixes an obvious typo. Please apply.

Bad semicolon in previous patch ([1/2]). Thanks to
Erik@harddisk-recovery.nl for spotting this.

update CPUfreq documentation, and move it to extra directory.

Update x86 drivers so that they can live with the new cpufreq governor
"userspace".

sparc64 doesn't need any change, and I'll send the corresponding patch
for ARM to Russell King.

The old /proc/sys/cpu/ - based interface is nothing else than a
cpufreq governor which decides what frequency to use within a policy
based on userspace input. So, convert this interface (and add a
corresponding sysfs file, /sys/devices/sys/cpu0/scaling_setspeed) to
become a cpufreq governor, and move it out to an extra module (which
partly explains the size of this patch).

- update the VIA Longhaul driver to use frequency table helpers and
  the advanced cpufreq driver registration interface.
- cleanup of the ACPI P-States cpufreq driver
- very small update of the core -- set 24API values ahead of the setpolicy
  call.