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

into kernel.bkbits.net:/home/davem/tg3-2.6

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

into kernel.bkbits.net:/home/davem/sparc-2.6

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

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

It can return with the lock held.

Found by the Stanford locking checker.
Signed-off-by: Andrew Morton <akpm@osdl.org>

into pobox.com:/spare/repo/net-drivers-2.6

Included is a patch for linux to add a PCI vendor/device ID for the
Radisys ENP-2611 board.  The ENP-2611 is a 64bit/66MHz PCI board which
hosts an Intel IXP2400 network processor, has three GigE interfaces,
runs linux and generally kicks ass.

(see http://www.radisys.com/oem_products/ds-page.cfm?productdatasheetsid=1147)
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Please review and apply the following patch if you find it agreeable.

This patch does not make any functional changes, but does improve
both performance and memory usage by rearranging structure elements.

The need for these changes became appearent during a code review of
the disassembly involving this structure. The memory footprint of this
structure is made smaller by grouping the byte fields next to each other.
The access of the list_head can be simplified by making it the first element
of the structure, thus avoiding a needless add-immediate without negatively
impacting any of the other accesses.
Signed-off-by: Linas Vepstas <linas@linas.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

This patch fixes a null-pointer dereference when hot-plug operations
are performed on a machine that has virtual-io devices in it.
Virtual i/o devices to not have pci bridges associated with them.
It also corrects an ordering problem during hotplug remove.

This patch was previously reviewed/tested by Linda Xie, the current
rpaphp maintainer.
Signed-off-by: Linas Vepstas <linas@linas.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

This patch adds some PCI Express register constants to <linux/pci.h>

For my device, setting the Max_Read_Request_Size value in the PCI
Express device control register makes a huge performance difference.
I wanted my driver code that does this to be a little more
self-documenting than:

	pci_read_config_word(mdev->pdev, cap + 8, &val);
	val = (val & ~(5 << 12)) | (5 << 12);

I went a little overboard and added all the basic device register
fields.  If desired I could go even further overboard and add the
link, slot and root registers as well.

This patch is based on Matthew Wilcox's patch for pciutils, corrected
for some PCI Express spec 1.0a changes.
Signed-off-by: Roland Dreier <roland@topspin.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Signed-off-by: Linda Xie lxie@us.ibm.com
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

 - fix possible prefetch-fault loop on NX page, based on suggestions
   from Jamie Lokier.

 - clean up nx feature dependencies

 - simplify detection of NX-violations when the kernel executes code

 - introduce pte_exec_kern() to simplify the NX logic

 - split the definitions out of pgtable-[23]level.h into
   pgtable-[23]level-defs.h, to enable the former to use generic
   pte functions from pgtable.h.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This patch, originally from Daniel Drake, replaces the meye driver
'wait_ms()' function with calls to the kernel provided 'msleep()'
function.
Signed-off-by: Daniel Drake <dsd@gentoo.org>
Signed-off-by: Stelian Pop <stelian@popies.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This updates the sonypi driver by:
 * fixing the power management handling, using the new device
   model PM scheme.

 * adds "VGN-" as a DMI search pattern for a Sony Vaio laptop.

Florian Lohoff reported the power management issue and tested the
patch.

Many users reported the DMI name issue, including Till Busch who
made a patch for dmi_scan.c.
Signed-off-by: Stelian Pop <stelian@popies.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

From the new Stanford locking checker
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

From the new Stanford locking checker
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Attached is a smallish patch for couple trivial sparse warnings in
allnoconfig build and more importantly an "excuses" text file explaining
why the rest have not been fixed.

Basically all of them (with the exception of the one in Andrews tree) need
some serious re-engineering.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The attached patch fixes the pushf under v86 to always set the IOPL field
to 3, as the Intel CPUs do.

It was in 2.4 for year, but somehow missed 2.6.  It comes from here
(although now a bit shorter):

http://lkml.org/lkml/2003/5/25/81Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Remove the last leftovers of the compatibility code for running the IRIX X
server.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Update my email address.  Add Manish Lachwani as maintainer for the MV-64340
driver which has already been accepted by Jeff Garzik.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

 - Remove support for NEC Eagle which is a system that was produced in
   almost zero numbers.
 - Sort out the MV-64340 code a little more.
 - Fix computation of interrupts masks for the upper 32 interrupts of the
   MV-64340.
 - Gas was producing bad break codes for MIPS32/MIPS64 processors for a while.
   Add a workaround to the kernel trap handler.
 - Rewrite RM7000 cache code in a way where it's hopefully easier to read.
 - Handle 64-bit pointers right for the BCM1250.
 - Try to share code for GT-64240 and MV-64340.
 - Fix RM9000 definition of _CACHE_CACHABLE_NONCOHERENT.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

CONFIG_DECSTATION is now CONFIG_MACH_DECSTATION ...
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

I recently got an ISA Phonejack card and quickly found it wouldn't work
with a stock 2.6.7 kernel. 

So found this bug in the ISAPNP part of the code, throughout the code it
use it's own data structure to reference it's base IO address, in the
ISAPNP code this was not being populated with any value hence it can't talk
to the card and so the driver fails.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

From: Heiko Carstens <heiko.carstens@de.ibm.com>
From: Andreas Herrmann <aherrman@de.ibm.com>
From: Maxim Shchetynin <maxim@de.ibm.com>

zfcp host adapter changes:
 - Exploit FC transport class and autoselect SCSI_FC_ATTRS for zfcp.
 - Fix acl download to zfcp controller.
 - Change message loglevels to make zfcp less noisy.
 - Don't wait for SBAL to finish for command aborts after a timeout
   and for logical unit or target resets.
 - Force reopen of port if link test failed.
 - Fix race between qdio_shutdown and do_QDIO.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

From: Ursula Braun-Krahl <braunu@de.ibm.com>
From: Frank Pavlic <pavlic@de.ibm.com>
From: Thomas Spatzier <tspat@de.ibm.com>
From: Peter Tiedemann <ptiedem@de.ibm.com>

s390 network driver changes:
 - ctc: replace snprintf by strlcpy.
 - lcs: change info text for lcs cards from "OSA2 card" to "OSA LCS card".
 - lcs: fix alignment of lcs_cmd structure to get multicast pings working.
 - lcs: first call in_dev_put then register multicast addresses.
 - netiucv: remove unused device timer and unused flags field.
 - netiucv: include interrupt type in pathid mismatch message.
 - qeth: don't start a new kernel thread for every new ip address.
 - qeth: fix IP assist command sequence numbers.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

From: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>

sclp console driver changes:
 - Correct handling of busy and not operational states.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

From: Horst Hummel <horst.hummel@de.ibm.com>
From: Carsten Otte <cotte@de.ibm.com>
From: Stefan Weinhuber <wein@de.ibm.com>

dasd device driver changes:
 - Fix calculation of number of idal words needed for a channel program.
 - Fix race in i/o termination after request timeout.
 - Fix race in state change interrupt handling.
 - Fix call to BLKPG ioctl in dasd_destroy_partitions.
 - Integrate irb into dasd request to avoid kmalloc in the interrupt handler.
 - Store build clock for error recovery requests.
 - Remove unused cpu variable from dasd_ext_handler.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

From: Utz Bacher <utz.bacher@de.ibm.com>
From: Cornelia Huck <cohuck@de.ibm.com>

Common i/o layer changes:
 - Consolidate store channel subsystem characteristics from its three
   users (css, cmf and qdio) to a single location.
 - Always use new stipd format and move creation of global path group
   to channel subsystem init function. Add dummy init_IRQ to setup.c
   and remove requestirq.c.
 - Remove bogus CHPID_LONGS define.
 - Add more magic to catch chpids coming online again without generating
   machine checks.
 - Fix check for unsolicited interrupts. Deferred cc=1 indicates a
   solicited interrupt.
 - Fix progress indication in qdio summary bytes to avoid loosing interrupts.
 - Rename console_device to console_devno to avoid naming conflict.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

From: Arnd Bergmann <arndb@de.ibm.com>
From: Christian Bornträger <cborntra@de.ibm.com>
From: Michael Holzheu <holzheu@de.ibm.com>
From: Martin Schwidefsky <schwidefsky@de.ibm.com>

s390 core changes:
 - Fix cpu_idle loop if /proc/sys/kernel/hz_timer is set.
 - Store correct trap indication on 64 bit for call to do_debugger_trap
   in the single stepped svc code.
 - Avoid the use of alloca in the debug feature.
 - Remove extraneous includes of linux/version.h.
 - Regenerate default configuration.
 - Mention eServer z890 in Kconfig help text.
 - Prevent gcc 3.4 from removing statically defined per cpu variables.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The radix tree functions __lookup and __lookup_tag uses (1 << shift) in
their index calculations.  On 64 bit systems the shift can be bigger than
32.  The shift of an integer by more than 32 bits evaluates to zero which
causes the lookup to fail.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The init function is writing somewhere using a physical address instead of
a virtual address.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This is ES7000 sub arch update.  The tiny patch below fixes a bug that
results in boot problem on all ES7000 partitions with numbers higher than
0.  Currently, target_cpus() returns incorrect cpumask value since it
indexes bits by the logical APIC ID, not by CPU number.  The patch affects
only ES7000 sub arch.  It was tested extensively on the system with
multiple partitions.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Include <asm/io.h> for ioremap() declaration.
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

- Update display->vrows when fbset changes certain fields in var (such as
  color depth) but without changing the display resolution.  This might
  produce concommitant changes in other fields such as and especially
  var->yres_virtual.

- Clear the FBINFO_MISC_MODECHANGEUSER flag _before_ issuing a
  notifier_call_chain().  Clearing it after the notifier_call_chain() will
  cause fbcon to go into an infinite loop.

- Sanitized failure path of set_con2fb_map().
Signed-off-by: Antonino Daplas <adaplas@pol.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

As we've discussed in another thread, below is a diff that will do a set_par()
as late as possible when there is KD_TEXT<->KD_GRAPHICS switch.  The set_par()
will be forced in fbcon_resize() instead.

Not sure if this has repercussions with the other drivers, but this patch
fixed the X nv driver hanging when switching to the console.  (I believe the
crash is actually caused by an early set_par() -- while in fbcon_blank. 
Removing the set_par in fbcon_blank fixed the hang but caused cursor sprite
and display corruption).
Signed-off-by: Antonino Daplas <adaplas@pol.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>