From: Manfred Spraul <manfred@colorfullife.com>

The attached patch adds a simple kmem_cache_alloc_node function: allocate
memory on a given node.  The function is intended for cpu bound structures.
 It's used for alloc_percpu and for the slab-internal per-cpu structures.
Jack Steiner reported a ~3% performance increase for AIM7 on a 64-way
Itanium 2.

Port maintainers: The patch could cause problems if CPU_UP_PREPARE is
called for a cpu on a node before the corresponding memory is attached
and/or if alloc_pages_node doesn't fall back to memory from another node if
there is no memory in the requested node.  I think noone does that, but I'm
not sure.

From: Manfred Spraul <manfred@colorfullife.com>

The slab allocator keeps track of the free objects in a slab with a linked
list of integers (typedef'ed to kmem_bufctl_t).  Right now unsigned int is
used for kmem_bufctl_t, i.e.  4 bytes per-object overhead.

The attached patch implements a per-arch definition of for this type:
Theoretically, unsigned short is sufficient for kmem_bufctl_t and this would
reduce the per-object overhead to 2 bytes.  But some archs cannot operate on
16-bit values efficiently, thus it's not possible to switch everyone to
ushort.

The chosen types are a result of dicussions with the various arch maintainers.

From: Manfred Spraul <manfred@colorfullife.com>

the attached patch switches the SLAB_HWCACHE_ALIGN alignment from the
compile time L1 cache line size to the runtime detected value for i386. 
x86-64 already uses the runtime detection.

From: Nick Piggin <nickpiggin@yahoo.com.au>

If the zone has a very small number of inactive pages, local variable
`ratio' can be huge and we do way too much scanning.  So much so that Ingo
hit an NMI watchdog expiry, although that was because the zone would have a
had a single refcount-zero page in it, and that logic recently got fixed up
via get_page_testone().

Nick's patch simply puts a sane-looking upper bound on the number of pages
which we'll scan in this round.


It fixes another failure case: if the inactive list becomes very small
compared to the size of the active list, active list scanning (and therefore
inactive list refilling) also becomes small.

This patch causes inactive list scanning to be keyed off the size of the
active+inactive lists.  It has the plus of hiding active and inactive
balancing implementation from the higher level scanning code.  It will
slightly change other aspects of scanning behaviour, but probably not
significantly.

Experimenting with various values of DENTRY_STORAGE

 dentry size      objs/slab   dentry size * objs/slab    inline string

   148               26               3848                   32
   152               26               3952                   36
   156               25               3900                   40
   160               24               4000                   44

We're currently at 160.  The patch fairly arbitrarily takes it down to 152, so
we can fit a 35-char name into the inline part of the dentry.

Also, go back to the old way of sizing d_iname so that any arch-specific
compiler-forced alignemnts are honoured.

Fix http://bugme.osdl.org/show_bug.cgi?id=2710.

When the user passed madvise a length of -1 through -4095, madvise blindly
rounds this up to 0 then "succeeds".

From: Brian Gerst <bgerst@didntduck.org>

Generate offsets for thread_info, cpuinfo_x86, and a few others instead of
hardcoding them.

It conflicts with the readq() I/O function.

From: Chris Wright <chrisw@osdl.org>

Add disable param to capabilities module.  Similar to the SELinux param for
disabling at boot time.  This allows vendors to ship single binary image with
capabilities compiled statically, and disable it if they provide another
security model compiled as module.

From: Ram Pai <linuxram@us.ibm.com>

Currently the readahead code tends to read one more page than it should with
seeky database-style loads.  This was to prevent bogus readahead triggering
when we step into the last page of the current window.

The patch removes that workaround and fixes up the suboptimal logic instead.


wrt the "rounding errors" mentioned in this patch, Ram provided the following
description:

  Say the i/o size is 20 pages.

  Our algorithm starts by a initial average i/o size of 'ra_pages/2' which
  is mostly say 16.

  Now every time we take a average, the 'average' progresses as follows
  (16+20)/2=18
  (18+20)/2=19
  (19+20)/2=19
  (19+20)/2=19.....
  and the rounding error makes it never touch 20



Benchmarking sitrep:

			IOZONE

	run on a nfs mounted filesystem:
	client machine 2proc, 733MHz, 2GB memory
	server machine 8proc, 700Mhz, 8GB memory

./iozone -c -t1 -s 4096m -r 128k

drivers/scsi/dpt_i2o.c: In function `adpt_queue':
drivers/scsi/dpt_i2o.c:442: warning: use of cast expressions as lvalues is deprecated
drivers/scsi/dpt_i2o.c: In function `adpt_scsi_register':
drivers/scsi/dpt_i2o.c:2213: warning: use of cast expressions as lvalues is deprecated

From: Arthur Othieno <a.othieno@bluewin.ch>

CONFIG_MAC_SERIAL (drivers/macintosh/macserial.c) is marked obsolete and
currently doesn't build.

benh says: "I though build got fixed recently ...  well, anyway, the driver is
indeed obsolete, there's a new one in drivers/serial now."

From: Matt Porter <mporter@kernel.crashing.org>

Fixes a warning and a printk format in OCP.

From: Olaf Hering <olh@suse.de>

Missing spin_unlock in the error path.

From: Stephen Rothwell <sfr@au1.ibm.com>

This patch stops the iseries_veth driver trying to send every packet to too
many logical partitions.  Consequently, the number of transmit errors falls to
(about) zero from a very large number.  This should also improve performance a
bit as the driver is no longer doing 31 extra skb_clone()s and skb_free()s for
each packet.

From: Paul Mackerras <paulus@samba.org>

At present on ppc32, if the kernel accesses a bad address and causes an
oops, or drops into the xmon debugger, we only have the contents of the
volatile registers available to print.  The reason is that we only save the
volatile registers on entry for a page fault.

This patch restructures the code a bit so that if do_page_fault()
determines that the page fault is caused by a bad kernel access, it returns
to the caller, which then saves the full register set into the exception
frame before calling bad_page_fault().  This way we get the full set of
registers printed in the oops message.

From: Paul Mackerras <paulus@samba.org>

This patch adds module tags for the swim3 (macintosh floppy) driver.

blk_run_page() is incorrectly using page->mapping, which makes it racy against
removal from swapcache.

Make block_sync_page() use page_mapping(), and remove bkl_run_page(), which
only had one caller.

Split the system_state state `SYSTEM_SHUTDOWN' into SYSTEM_HALT,
SYSTEM_POWER_OFF and SYSTEM_RESTART and export system_state to modules.

This allows driver shutdown routines to know why they are being shutdown.  The
IDE subsystem wants this so that it knows to not spin the disks down across a
reboot.

into ppc970.osdl.org:/home/torvalds/v2.6/linux

From: Ivan Kokshaysky <ink@jurassic.park.msu.ru>

This allows building the math-emu code as a module only when
CONFIG_SMP is not set. The fp trap handler cannot be preempted
on a single-CPU (as CONFIG_PREEMPT is not going to be supported
on alpha), so the module can be safely unloaded at any time.

Avoid warnings about truncating them when building the
name database.

into ppc970.osdl.org:/home/torvalds/v2.6/linux

From: Chris Wedgwood <cw@f00f.org>

Seperate function, cruft removed and old_hwif
renamed to something less confusing.

From: Yoshinori Sato <ysato@users.sourceforge.jp>

With minor fixes from me

reiserfs_file_write unlocks the pages it operated on before updating
i_size.  This can lead to races with writepage, who checks i_size when
deciding how much of the file to zero out.

This patch also replaces SetPageReferenced with mark_page_accessed() in
reiserfs_file_write

This was verified to fix the BitKeeper data corruption problems that
Steven Cole has been debugging, where concurrent writes to a file and
writebacks to disk would cause zeroes in the file when CONFIG_PREEMPT
was enabled.

- clear hwif->hw in setup-pci.c before using it

- fix arch/arm/Kconfig to allow IDE only on platforms supporting it

- introduce IDE_ARCH_OBSOLETE_INIT and ide_default_io_ctl() so
  we can use generic ide_init_hwif_ports() and kill no longer needed
  <asm-arm/arch-*/ide.h> (leave broken lh7a40x and sa1100 versions)

Cross-compile tested on ARM.

Add drivers/ide/arm/ide_arm.c for simple default IDE interfaces
and clean obsolete ide_init_default_hwifs() implementations
in asm-arm/arch-{cl7500,rpc,shark}/ide.h and asm-arm26/ide.h.

This allows us to kill ide_init_default_hwifs() completely
in the next patch (because lh7a40x and sa1100 are broken).

Cross-compile tested on ARM.

* Convert all netdev->priv references to the fancy new netdev_priv().

* Reads/writes from/to eeprom using ethtool weren't working right
  on big-endian.  Now they are.

* Changed mapping on Rx skb to bi-directional.  skb->data holds both the
  RFD structure and the packet data, and the RFD is read/written by
  HW.  Issue found on Xscale HW that doesn't handle cache syncs auto-
  matically.  Other changes in patch are whitespace/spelling.

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into ppc970.osdl.org:/home/torvalds/v2.6/linux

Bryan Rosenburg pointed out that the definition of PT_FPSCR in
include/asm-ppc64/ptrace.h is wrong.  The patch below fixes it.

From: Markus Lidel <Markus.Lidel@shadowconnect.com>

I have made a mistake in the kernel header i2o-dev.h.  All structures there
begin with the "iop" first, but in my structure the order is reversed.