Fix up some misc warnings: include module.h in cache.c since it uses
EXPORT_SYMBOL(), fix printk format arg in sn_serial.c.

The note_interrupt() code is racy for per-CPU irqs and causes severe
performance problems due to cache-line bouncing on large SMPs.  Just don't
call it for that case.

	via /proc/efi/fpswa, rather than printing it at boot time.

Periodically check for outstanding MCA or INIT records and pass them to
user space salinfo.

Correct the "did we recover from MCA test" and move it up a level to
simplify interaction with debuggers.

Patches from Ben Woodward to calculate irq_safe once and pass it around.

[PATCH] ia64: mca.c cleanup - Reorder to remove the need for forward declarations and to consolidate related code

[PATCH] ia64: mca.c cleanup - Delete all record printing code, moved to salinfo_decode in user space

Port the ia64 mca.c clean up patches from 2.4.25-pre8 to 2.6.2-rc2.

The following 6 patches do :-

1 Avoid deadlock when using printk() for MCA and INIT records.
2 Delete all record printing code, moved to salinfo_decode in user space.
3 Mark variables and functions static where possible.
4 Delete dead variables and functions.
5 Reorder to remove the need for forward declarations and to consolidate
  related code.
6 Bjorn's printk cleanup.

Altogether they shrink mca.c from 2432 to 1339 lines and make it much
more readable.

The only functional change is the removal of any attempt to print the
CMC/CPE/MCA/INIT record contents in the kernel plus the addition of an
info printk to ia64_mca_check_errors(), to match 2.4.  Now we just get
one line to say that a record has been detected, except for MCA which
prints nothing at all.

   I've been doing some performance tuning and adding some functionality
to sba_iommu for zx1/sx1000 chipsets.  This adds:

      * Long overdue consistent_dma_mask support
      * Long overdue ability to do large mappings in the iommu
      * Tightened spinlock usage for better performance/scalability
      * Added branch prediction hints for some of the performance paths
      * Added explicit data prefetching to some performance paths -
        perfmon shows roughly a 20% decrease in L3 misses in the bitmap
        search code
      * Increased delayed resource freeing depth and added a separate
        lock per ioc to avoid contention
      * Added code to free up queued pdir entries should we be unable to
        find space for new ones (not that I've ever seen the pdir
        anywhere close to full)
      * Finished cleaning out the hint support code, Grant is
        maintaining this separately for now
      * Added option to control bypass of sg mappings separately from
        single/coherent mappings

Much like the swiotlb, sba_iommu allows devices capable of 64bit
addressing to bypass the iommu and DMA directly to/from memory.  Using a
worst case scenario test (64bit bypass disabled, all DMA mapped through
the iommu), I saw a 60% increase in sequential block input throughput
using bonnie++ on a large RAID0 MD array.  In fact, this patch provides
the best bonnie++ performance with bypass disabled.  This is likely due
to benefits seen from coalescing the scatterlist, allowing better disk
streaming.  I assume that network performance will likely be limited by
mapping latency, so I added the last bullet item to allow sg mappings to
get the benefit of coalescing while keeping a low latency path for
single and coherent mappings.  If anyone is setup for network
benchmarks, I'd be interested in a before and after with this patch.

This patch fixes a corner-case of ld.a emulation.  ld.a should be
emulated such that it always loads a misaligned value and clears the
corresponding ALAT entry.  The old emulation was correct for the case
where ld.a was followed by ld.c/chk.a (since ALAT entry was cleared),
but wrong for the case where it _wasn't_ followed by ld.c/chk.a.  In
that case, the misaligned value wasn't read from memory, as it should
have been.

The S_IFxxxx macros aren't bits to be tested, they
are values of the S_IFMT field.

This patch fixes the GPM cursor corruption people where seeing.

With 32 bit sector_t the block device size _in bytes_ is also cut to
32 bit in __set_size when the block device is mount (a filesystem
mounted). The argument should be cast to loff_t before expanding the
sector count to a byte count and calling i_size_write.

[Christophe Saout]

user program inclusion.

It's still wrong to include kernel headers from
user programs. Oh, well.

Ths "fix duplication of DMA {black,white}list in icside.c" patch broke it.

Noticed by Geert Uytterhoeven <geert@linux-m68k.org>.

From John McKell <mckellj@iomega.com>:

  This patch for scsi/sr.c and cdrom/cdrom.c persuades
  /proc/sys/dev/cdrom/info to report that connecting a drive via USB
  rather than ATAPI in fact does not make it able to write DVD-R.

  Without this patch, when sr0 and hdd are the same type of device
  connected via USB and ATAPI respectively, I see:

	$ cat /proc/sys/dev/cdrom/info:
	drive name:             sr1     sr0     hdd
	...
	Can write CD-R:         1       0       0
	Can write CD-RW:        1       0       0
	Can read DVD:           0       0       0
	Can write DVD-R:        1       1       0
	Can write DVD-RAM:      0       0       0
	Can read MRW:           0               0               0
	Can write MRW:          0               0               0

  With this patch applied, instead I see:

	$ cat /proc/sys/dev/cdrom/info:
	drive name:             sr1     sr0     hdd
	...
	Can write CD-R:         1       0       0
	Can write CD-RW:        1       0       0
	Can read DVD:           0       0       0
	Can write DVD-R:        0       0       0
	Can write DVD-RAM:      0       0       0
	Can read MRW:           0       0       0
	Can write MRW:          0       0       0

  The sr1 device in particular is an ordinary CD-RW that in fact cannot
  write DVD-R.

  While messing with this code, I also thought to tweak the /proc
  formatting to align the tabbed columns.

Here's a new version of the patch according to our discussion.

This allows the platform to add its own platform files to the
sysfs PCI device node through a "pcibios_add_platform_entries()"
call.

I added an empty pcibios_add_platform_entries() inline to all
archs that apparently have PCI and implemented the OF stuff for
ppc and ppc64. The name is still "devspec" for now, I doubt
it will conflict and that's consistent with our use of the
"devspec" name in other places.

From: Nathan Scott <nathans@sgi.com>,
      Andreas Gruenbacher <agruen@suse.de>

XFS regression tests tripped a couple of problems with the recent xattr fix.

When a size of 0 is passed in, the getxattr and listxattr syscalls return the
size that would be required for the value or list of names, without actually
returning the value.  The previous patch accidentally removed this test, and
so querying the required size broke.

From: "Art Haas" <ahaas@airmail.net>

The following patch removes the duplicated sysctl entries for 'soft-power'
and 'unaligned-trap' for HPPA users.  Immediately before these lines these
two entries exist in a C99-style format.

into home.osdl.org:/home/torvalds/v2.5/linux

into home.osdl.org:/home/torvalds/v2.5/linux

<Patch from Stephen Hemminger>
o [CORRECT] Better handling of shared IRQs in w83977af_ir driver.

<Patch from Stephen Hemminger>
o [CORRECT] Better handling of shared IRQs in via-ircc driver.

<Patch from Stephen Hemminger>
o [CORRECT] Better handling of shared IRQs in smsc-ircc2 driver.

<Patch from Stephen Hemminger>
o [CORRECT] Better handling of shared IRQs in ali-ircc driver.

<Patch from Stephen Hemminger>
o [CORRECT] Better handling of shared IRQs in nsc-ircc driver.

<Patch from Stephen Hemminger>
o [FEATURE] remove unused code : device notifier handler.

o [FEATURE] Don't drop IrLAP connection is we *just* received an
	incomming IrLMP connection request.

<Original patch from Stephen Hemminger>
o [CORRECT] Always initialise Ultra packet/header size.
o [CORRECT] Don't allow Ultra send on unbound sockets if no
	dest address is given.
o [FEATURE] Properly support Ultra sendto on unbound sockets.