[PKT_SCHED]: Fix dsmark to count ignored indices while walking

Unused indices which are ignored while walking must still
be counted to avoid dumping the same index twice.

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

[IPV4]: Fix BUG() in 2.6.x, udp_poll(), fragments + CONFIG_HIGHMEM

Steven Hand <Steven.Hand@cl.cam.ac.uk> wrote:
>
> Reconstructed forward trace:
>
>   net/ipv4/udp.c:1334   spin_lock_irq()
>   net/ipv4/udp.c:1336   udp_checksum_complete()
> net/core/skbuff.c:1069   skb_shinfo(skb)->nr_frags > 1
> net/core/skbuff.c:1086   kunmap_skb_frag()
> net/core/skbuff.h:1087   local_bh_enable()
> kernel/softirq.c:0140   WARN_ON(irqs_disabled());

The receive queue lock is never taken in IRQs (and should never be) so
we can simply substitute bh for irq.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

[NETFILTER]: Fix deadlock with ip_queue and tcp local input path.

When we have ip_queue being used from LOCAL_IN, then we end up with a
situation where the verdicts coming back from userspace traverse the TCP
input path from syscall context.  While this seems to work most of the
time, there's an ugly deadlock:

syscall context is interrupted by the timer interrupt.  When the timer
interrupt leaves, the timer softirq get's scheduled and calls
tcp_delack_timer() and alike.  They themselves do bh_lock_sock(sk),
which is already held from somewhere else -> boom.

I've now tested the suggested solution by Patrick McHardy and Herbert Xu to
simply use local_bh_{en,dis}able().

Signed-off-by: Harald Welte <laforge@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge ... /pub/scm/linux/kernel/git/davem/tg3-2.6

Merge of /home/davem/src/GIT/tg3-2.6/

[NET]: Use %lx for netdev->features sysfs formatting.

Signed-off-by: David S. Miller <davem@davemloft.net>

[IPV6]: Clear up user copy warning in flowlabel code.

We are intentionally ignoring the copy_to_user() value,
make it clear to the compiler too.

Noted by Jeff Garzik.

Signed-off-by: David S. Miller <davem@davemloft.net>

[NET]: Add ethtool support for NETIF_F_HW_CSUM.

Signed-off-by: Jon Mason <jdmason@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[IPV4]: Kill MULTIPATHHOLDROUTE flag.

It cannot work properly, so just ignore it in drr
and rr multipath algorithms just like the random
multipath algorithm does.

Suggested by Herbert Xu.

Signed-off by: Pravin B. Shelar <pravins@calsoftinc.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[NET]: Add is_multicast_ether_addr() in include/linux/etherdevice.h

This patch adds is_multicast_ether_addr() to go along with
is_valid_ether_addr() and friends. It then changes
is_valid_ether_addr() to use the new macro, and fixes up the comment
on that function to move implementation details out of the API doco.

Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

[IPV4]: Primary and secondary addresses

Add an option to make secondary IP addresses get promoted
when primary IP addresses are removed from the device.
It defaults to off to preserve existing behavior.

Signed-off-by: Harald Welte <laforge@gnumonks.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Automatic merge of 'misc-fixes' branch from

rsync://rsync.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev

[TG3]: Fix bug in tg3_load_firmware_cpu

Add tg3_nvram_lock() and tg3_nvram_unlock() calls around tg3_halt_cpu().
It is possible that the bootcode may be loading code from nvram during
this call and stopping the cpu without getting the lock may cause
uncompleted nvram data to be left in the nvram data register. Subsequent
calls to read/write nvram data will fail.

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[TG3]: Add interrupt test

This test uses the previously added tg3_test_interrupt() to perform the
test.

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[TG3]: Add loopback test

The test will loopback one packet in MAC loopback mode and verify the
packet data.

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[TG3]: Add memory test

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[TG3]: Add register test

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[TG3]: Add parameter to tg3_halt

Add a reset kind parameter to tg3_halt() so that the RESET_KIND_SUSPEND
parameter can be passed to tg3_halt() before doing offline tests.

All other calls to tg3_halt() will use the RESET_KIND_SHUTDOWN
parameter.

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[TG3]: Add link test

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[TG3]: Add nvram test

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[TG3]: Add basic selftest infrastructure

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

[BRIDGE]: receive path optimization

This improves the bridge local receive path by avoiding going
through another softirq.  The bridge receive path is already being called
from a netif_receive_skb() there is no point in going through another
receiveq round trip.

Recursion is limited because bridge can never be a port of a bridge
so handle_bridge() always returns.

Signed-off-by: David S. Miller <davem@davemloft.net>

[BRIDGE]: prevent bad forwarding table updates

Avoid poisoning of the bridge forwarding table by frames that have been
dropped by filtering. This prevents spoofed source addresses on hostile
side of bridge from causing packet leakage, a small but possible security
risk.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

[BRIDGE]: set features based on enslaved devices

Make features of the bridge pseudo-device be a subset of the underlying
devices.  Motivated by Xen and others who use bridging to do failover.

Signed-off-by: Catalin BOIE <catab at umrella.ro>
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

[BRIDGE]: make dev->features unsigned

The features field in netdevice is really a bitmask, and bitmask's should
be unsigned.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

[BRIDGE]: features change notification

Resend of earlier patch (no changes) from Catalin used to provide
device feature change notification.

Signed-off-by: Catalin BOIE <catab at umbrella.ro>
Acked-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

libata: bump version

Merge ... /spare/repo/netdev-2.6 branch use-after-unmap

Automatic merge of 'for-linus' branch from

rsync://rsync.kernel.org/pub/scm/linux/kernel/git/dtor/input

Input: Fix fast scrolling scancodes in atkbd.c

Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

[PATCH] ppc32: i8259 PIC should not be initialized if PCI is not configured

Trying to initialize the i8259 PIC will not work if CONFIG_PCI is not
enabled.  The kernel hangs if the initialization is tried.

Signed-off-by: Kumar Gala <kumar.gala@freescale.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ppc32: Add soft reset to MPC834x

This change allows mpc83xx_restart to issue a software reset.

Signed-off-by: Randy Vinson <rvinson@mvista.com>
Signed-off-by: Kumar Gala <kumar.gala@freescale.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ppc32: MPC834x BCSR_SIZE too small for use in a BAT.

The call to io_block_mapping was creating an invalid BAT entry because the
value of BCSR_SIZE (32K) is too small to be used in a BAT (128K min).  This
change removes the io_block_mapping call since these registers can easily
be mapped using ioremap at the point of use.

Signed-off-by: Randy Vinson <rvinson@mvista.com>
Signed-off-by: Kumar Gala <kumar.gala@freescale.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ppc32: Simplified load string emulation error checking

The error checking for emulation of load string instructions was overly
generous and would cause certain valid forms of the instructions to be
treated as illegal.  We drop the range checking since the architecture
allows this to be boundedly undefined.  Tests on CPUs that support these
instructions appear not do cause illegal instruction traps on range errors
and just allow the execution to occur.

Thanks to Kim Phillips for debugging this and figuring out what real HW was
doing.

Signed-off-by: Kumar Gala <kumar.gala@freescale.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ppc32: Fix building MPC8555 CDS when CONFIG_PCI is disabled

The patch that introduced support for the VIA chipset broke building if
CONFIG_PCI is disabled.

Signed-off-by: Kumar Gala <kumar.gala@freescale.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ppc32: Support for 82xx PQII on-chip PCI bridge

This patch adds on-chip PCI bridge support for the PQ2 family.  The
incomplete existent code is updated with interrupt handling stuff and
board-specific bits for 8272ADS and PQ2FADS; the related files were renamed
(from m8260_pci to m82xx_pci) to be of more generic fashion.  This is
tested with 8266ADS and 8272ADS, should work on PQ2FADS as well.

Signed-off-by: Vitaly Bordug <vbordug@ru.mvista.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ppc32: Add VIA IDE support to MPC8555 CDS platform

Add support for the VIA IDE controller that exists on the MPC8555 CDS
system.  Updated the config for the system to enable support by default.

Signed-off-by: Scott Hall <shall@mvista.com>
Signed-off-by: Kumar Gala <kumar.gala@freescale.com>
Cc: Bartlomiej Zolnierkiewicz <B.Zolnierkiewicz@elka.pw.edu.pl>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ppc32: Fix some minor issues related to FSL Book-E KGDB support

Some debug registers needed to be initialized early on to allow proper
support for KGDB.  Additionally, we need to setup the
ppc.md_early_serial_map function pointer on boards that have serial support
for KGDB.

Signed-off-by: Kumar Gala <kumar.gala@freescale.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ppc32: Fix uImage make target to report success correctly

The existing make rule when building a uImage would check to see if the
image file existed to report 'is ready' or 'not made'.  However make
appeared to compute the file list before the rule was executed.

Signed-off-by: Chris Clark <cpclark@xmission.com>
Signed-off-by: Kumar Gala <kumar.gala@freescale.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: remove jail mode + other leftovers

This var is currently useless, as it's apparent from reading the code. Until
2.6.11 it was used in some code related to jail mode, in the same proc.:

        if(jail){
while(!reading) sched_yield();
}

jail mode has been dropped, together with that use, so let's finish dropping
this.

Also, remove some other useless definitions I met.

Acked-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: fixlet for arch_prctl_skas

Fix it a bit (after some cross checking with "man arch_prctl"). There were:
*) typos FS/GS and back
*) FS in place of FS_BASE (and the same for GS)
*) the procedure used put_user on &addr, where addr was the parameter (i.e.
changed its param with put_user, completely useless) rather than interpreting
addr as a pointer, as requested in this case (see the man page).

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] irq code: Add coherence test for PREEMPT_ACTIVE

After porting this fixlet to UML:

http://linux.bkbits.net:8080/linux-2.5/cset@41791ab52lfMuF2i3V-eTIGRBbDYKQ

, I've also added a warning which should refuse compilation with insane values
for PREEMPT_ACTIVE...  maybe we should simply move PREEMPT_ACTIVE out of
architectures using GENERIC_IRQS.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] sh: PREEMPT_ACTIVE fix

Port Paulus's ppc64 fix to sh:

  When the generic IRQ stuff went in, it seems that HARDIRQ_BITS got bumped
  from 9 (for ppc64) up to 12.  Consequently, the PREEMPT_ACTIVE bit is now
  within HARDIRQ_MASK, and I get in_interrupt() falsely returning true when
  PREEMPT_ACTIVE is set, and thus a BUG_ON tripping in arch/ppc64/mm/tlb.c.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: fix PREEMPT_ACTIVE

This is a continuation for UML of:

http://linux.bkbits.net:8080/linux-2.5/cset@41791ab52lfMuF2i3V-eTIGRBbDYKQ

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: stack dump fix

Copy (and adapt) to UML the stack code dumper used in i386 when
CONFIG_FRAME_POINTER is enabled.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: split CONFIG_FRAME_POINTER from DEBUG_INFO

Until now, FRAME_POINTER was set = DEBUG_INFO for UML.  Change it to be the
default way, so that it can be enabled alone (for instance to get better
backtraces on crashes).  The call-trace dumper which uses the frame pointer is
not yet in, I'm going to introduce it in a separate patch.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: add MOD_LICENSE to random driver

Add a MODULE_LICENSE("GPL") to the driver, remove some unused macros and add
the GPL license (it's GPL-licensed anyway since it's a GPL-derivative, apart
that Jeff Dike releases GPL software, in case anybody is wondering).

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: add modversions support

Actually, the real support was added by some earlier patches.  Now we simply
re-enable the config.  option.  I've actually tested it and it works well.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: remove unused code

This removes two now unused files and a couple of unused functions.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: fix segfault on exit with CONFIG_GCOV

We need to disable signals on exit in all cases, not just when rebooting.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: single-space a help message

One of the ubd driver help strings was bust.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: remove 2_5compat.h

Remove old useless header that was used in Ye Olde Times during 2.4->2.5
porting to abstract differences.  It's definitions are no more used anyway, so
let's finally kill it.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: turn off kmalloc always on a fatal signal

We should turn off kmalloc when getting a fatal signal regardless of the mode
we're in.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] uml: fix a couple of warnings

Eliminate an unused variable warning in ptrace.c and a size mismatch warning
by adding a cast to __pa.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] dvb: Small cleanup

Miscellaneous cleanup

Signed-off-by: Manu Abraham <manu@kromtek.com>
Cc: Johannes Stezenbach <js@linuxtv.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] dvb: Fix 22k tone control

Signed-off-by: Manu Abraham <manu@kromtek.com>
Cc: Johannes Stezenbach <js@linuxtv.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] dvb: Fix LNB power switching

Signed-off-by: Manu Abraham <manu@kromtek.com>
Cc: Johannes Stezenbach <js@linuxtv.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] dvb: Remove unnecessary casts

Cleanup unnecessary (and undesirable) casts, demodulator_priv is already a
void*.  Suggestion from Andrew Morton

Signed-off-by: Manu Abraham <manu@kromtek.com>
Cc: Johannes Stezenbach <js@linuxtv.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] dvb: Fix Mini DiSEqC bug

The bug was visible as a warning with gcc-3.4.4 (prerelease)

Message:
drivers/media/dvb/bt8xx/dst.c:1349: warning: initialization from
incompatible pointer type.

Signed-off-by: Manu Abraham <manu@kromtek.com>
Cc: Johannes Stezenbach <js@linuxtv.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] input: Fix fast scrolling scancodes in atkbd.c

Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] intelfb section fix

On Nov 16 2004 a change to intelfbdrv.c was commited (as part of 0.9.2 it
looks like) that added __initdata to all of the module param variables that
seems to create the opportunity for an oops.

I've recently been chasing an OOPS
(http://marc.theaimsgroup.com/?l=linux-kernel&m=111552250920370&w=2) I
created by reading every file on the /sys file system and I've traced it
back to this code in the intelfbdrv.  Though I had root privs in my initial
problem report, it turns out they are un-necessary to generate the oops -
all you've got to do is "cat /sys/module/intelfb/parameters/mode" enough
times and eventually it will oops.

This is because sysfs automatically exports all module_param declarations
to the sysfs file system..  which means those variables can be dynamically
evaluated at any later time, which of course means marking them __initdata
is a bad idea ;)..  when they happen to be char *'s it is an especially bad
idea ;).

Applying the patch below clears up the OOPS for me.

Signed-off-by: Patrick McManus <mcmanus@ducksong.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] fix jumpy mouse cursor on console

Do not send empty events to gpm.  (Keyboards are assumed to have scroll
wheel these days, that makes them part-mouse.  That means typing on
keyboard generates empty mouse events).

From: Dmitry Torokhov <dtor_core@ameritech.net>
Signed-off-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] swsusp: ahd_dv_0 can't be stopped

This driver wants to set PF_NOFREEZE.

Cc: James Bottomley <James.Bottomley@steeleye.com>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ppc32: Fix cpufreq vs. sleep issue

Recent kernels occasionally trigger a PMU timeout on some mac laptops,
typically on wakeup from sleep.  This seem to be caused by either a too big
latency caused by the cpufreq switch on wakeup from sleep or by an
interrupt beeing lost due to the reset of the interrupt controller done
during wakeup.

This patch makes that code more robust by stopping PMU auto poll activity
around cpufreq changes on machines that use the PMU for such changes (long
latency switching involving a CPU hard reset and flush of all caches) and
by removing the reset of the open pic interrupt controller on wakeup (that
can cause the loss of an interrupt and Darwin doesn't do it, so it must not
be necessary).

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] x86: fix smp_num_siblings on buggy BIOSes

This fixes 'smp_num_siblings' value on the systems with a buggy bios,
which sets number of siblings to '2' even when HT is disabled.  (more
details are at http://bugzilla.kernel.org/show_bug.cgi?id=4359)

I am planning to do more cleanup in this area (like moving smp_num_siblings
to per cpuinfo) shortly.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] drop note_interrupt() for per-CPU for proper scaling

The "unhandled interrupts" catcher, note_interrupt(), increments a global
desc->irq_count and grossly damages scaling of very large systems, e.g.,
>192p ia64 Altix, because of this highly contented cacheline, especially
for timer interrupts.  384p is severely crippled, and 512p is unuseable.

All calls to note_interrupt() can be disabled by booting with "noirqdebug",
but this disables the useful interrupt checking for all interrupts.

I propose eliminating note_interrupt() for all per-CPU interrupts.  This
was the behavior of linux-2.6.10 and earlier, but in 2.6.11 a code
restructuring added a call to note_interrupt() for per-CPU interrupts.
Besides, note_interrupt() is a bit racy for concurrent CPU calls anyway, as
the desc->irq_count++ increment isn't atomic (which, if done, would make
scaling even worse).

Signed-off-by: John Hawkes <hawkes@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] arch/i386/kernel/cpu/intel_cacheinfo.c: section fix

num_cache_leaves is used in __devexit cache_remove_dev() and can therefore
not be __devinit.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ppc64 iSeries: make virtual DVD-RAMs writable again

It appears that another test has been added in the Uniform CDROM layer that
must be passed before a DVD-RAM is considered writeable.  This patch
implements an emulation of the needed packet command for the viocd driver.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ppc32: Fix Alsa PowerMac driver on old machines

The g5 support code broke some earlier models unfortunately as those bail
out early from the detect function, before the point where I added the code
to locate the PCI device for use with DMA allocations.

This patch fixes it.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ppc32: small cpufreq update

This patch adds a missing attribute to pmac cpufreq so that
"scaling_available_frequencies" works.  It also cleans up the duplicate
definitions for low and high speed constants.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: John Clemens <clemej@alum.rpi.edu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] x86_64: signal.c build fix

For unspecified reasons, arch/x86_64/kernel/signal.c apparently needs
ia32_unistd.h.

Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] Fix up pwc driver compilation.

The neutering of the pwc driver was incomplete. It still references
some now-dead files..

Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Input: synaptics - reduce verboseness of synaptics driver - there
       is no reason one driver should take 10 lines in dmesg.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input: yet another model that does not play nicely when i8042 is
       put in MUX mode - Fujitsu Lifebook S6230

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input: automatically disable MUX mode on Toshiba Satellite P10
       because it interferes with ALPS touchpad detection and
       causes horrible death on reboot. Since P10 does not have
       external PS/2 ports MUX mode does not have any advantages
       over legacy mode anyway.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input: gunze - fix out-of-bound array access reported by Adrian Bunk.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input: Fix button mapping in joydev - BTN_TRIGGER was being
       mapped twice, resulting in it being the last (instead
       of first) button on a joystick.

Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input: Workaround for Sunrex K8561 IR Keyboard/Mouse. The mouse
       sends an incorrect ID and wasn't recognized.

Reported-by: Stefan Seyfried <seife@suse.de>
Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input: Avoid double unregistering of i8042 PnP driver. This can happen
       when no i8042 controller (not PnP, not legacy) is present.

From: Kurt Garloff <garloff@suse.de>
Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input: Tone down the severity of a printk() in i386/ia64 arch code
       for i386, it's printed on many machines and usually is not
       a cause for worry.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>

Input: Only write the CTR in i8042 resume function. Reading it is
       wrong, since it may (will) contain nonsensical data.

Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input: Remove (now) unused variable in i8042.c

Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input: Add a missing KERN_INFO message designation, fix behavior
       when only a keyboard part of the controller is detected.

Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input:
This patch adds dummy gameport_register_port, gameport_unregister_port
and gameport_set_phys functions to gameport.h for the case when a driver
can't use gameport.

This fixes the compilation of some OSS drivers with GAMEPORT=n without
the need to #if inside every single driver.

This patch also removes the non-working and now obsolete SOUND_GAMEPORT.

This patch is also an alternative solution for ALSA drivers with similar
problems (but #if's inside the drivers might have the advantage of
saving some more bytes of gameport is not available).

The only user-visible change is that for GAMEPORT=m the affected OSS
drivers are now allowed to be built statically (but they won't have
gameport support).

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

Input: Fix a warning in psmouse-base.c

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>

[PATCH] drivers/scsi/ahci: add #include req'd for the DMA_{64,32}BIT_MASK constants

The previous patch did not compile cleanly on all architectures so
here's a fixed one which #includes <linux/dma-mapping.h>.

Use the DMA_{64,32}BIT_MASK constants from dma-mapping.h when calling
pci_set_dma_mask() or pci_set_consistent_dma_mask()
This patch includes dma-mapping.h explicitly because it caused errors
on some architectures otherwise.
See http://marc.theaimsgroup.com/?t=108001993000001&r=1&w=2 for details

Signed-off-by: Tobias Klauser <tklauser@nuerscht.ch>
Signed-off-by: Domen Puncer <domen@coderock.org>
diff -puN drivers/scsi/ahci.c~dma_mask-drivers_scsi_ahci drivers/scsi/ahci.c

[PATCH] drivers/scsi/sata_vsc: add #include req'd for DMA_32BIT_MASK constant

The previous patch did not compile cleanly on all architectures so
here's a fixed one which #includes <linux/dma-mapping.h>.

Use the DMA_{64,32}BIT_MASK constants from dma-mapping.h when calling
pci_set_dma_mask() or pci_set_consistent_dma_mask()
This patch includes dma-mapping.h explicitly because it caused errors
on some architectures otherwise.
See http://marc.theaimsgroup.com/?t=108001993000001&r=1&w=2 for details

Signed-off-by: Tobias Klauser <tklauser@nuerscht.ch>
Signed-off-by: Domen Puncer <domen@coderock.org>
diff -puN drivers/scsi/sata_vsc.c~dma_mask-drivers_scsi_sata_vsc drivers/scsi/sata_vsc.c

Merge ... /pub/scm/linux/kernel/git/hch/xfs-2.6

[PATCH] Note on ACPI build fix

Even after the previous fix you can still set CONFIG_ACPI_BOOT
indirectly even without CONFIG_ACPI by choosing CONFIG_PCI and
CONFIG_PCI_MMCONFIG.

That doesn't build very well either.

This makes PCI_MMCONFIG depend on ACPI, fixing that hole.

[ I guess in theory Kconfig could follow the whole chain of dependencies
  for things that get selected, but that sounds insanely complicated, so
  we'll just fix up these things by hand.  --Linus ]

Signed-off-by: Alexander Nyberg <alexn@telia.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] ACPI build fix

Fix 2.6.12 CONFIG_ACPI=n build regression.
CONFIG_ACPI_BOOT shall be set only if CONFIG_ACPI.

Signed-off-by: Len Brown <len.brown@intel.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] Fixup VIA IRQ quirk

quirk_via_irqpic can't be __devinit for swsuspend

Signed-off-by: Alexander Nyberg <alexn@telia.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] VIA IRQ quirk

Delete quirk_via_bridge(), restore quirk_via_irqpic() -- but now
improved to be invoked upon device ENABLE, and now only for VIA devices
-- not all devices behind VIA bridges.

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Len Brown <len.brown@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] cpuset exit NULL dereference fix

There is a race in the kernel cpuset code, between the code
to handle notify_on_release, and the code to remove a cpuset.
The notify_on_release code can end up trying to access a
cpuset that has been removed.  In the most common case, this
causes a NULL pointer dereference from the routine cpuset_path.
However all manner of bad things are possible, in theory at least.

The existing code decrements the cpuset use count, and if the
count goes to zero, processes the notify_on_release request,
if appropriate.  However, once the count goes to zero, unless we
are holding the global cpuset_sem semaphore, there is nothing to
stop another task from immediately removing the cpuset entirely,
and recycling its memory.

The obvious fix would be to always hold the cpuset_sem
semaphore while decrementing the use count and dealing with
notify_on_release.  However we don't want to force a global
semaphore into the mainline task exit path, as that might create
a scaling problem.

The actual fix is almost as easy - since this is only an issue
for cpusets using notify_on_release, which the top level big
cpusets don't normally need to use, only take the cpuset_sem
for cpusets using notify_on_release.

This code has been run for hours without a hiccup, while running
a cpuset create/destroy stress test that could crash the existing
kernel in seconds.  This patch applies to the current -linus
git kernel.

Signed-off-by: Paul Jackson <pj@sgi.com>
Acked-by: Simon Derr <simon.derr@bull.net>
Acked-by: Dinakar Guniguntala <dino@in.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

[PATCH] remove non-cleanroom pwc driver compression

The original pwc author raised some questions about the reverse
engineering of the decompressor algorithms used in the pwc driver.
Having done some detailed investigation it appears those concerns that
clean room policy was not followed are reasonable.  I've also had a
friendly discussion with Philips to ask their view on this.

This removes the problem items of code which reduces the pwc
functionality in the kernel a little but leaves all the framework for
setup that will be needed for decompressors in user space (where they
eventually belong).  This change set is designed to be the minimal risk
change set given that 2.6.12 is hopefully close to hand, with a view to
merging the much updated pwc code in 2.6.13 series kernels.

Someone else can then redo the decompressors properly (clean room) in
user space.

Note that while its easy to say that it should have been caught earlier,
but the violation was really only obvious to someone who had access to
both the proprietary source and the 'GPL' source.

ide-cd: revert DMA mask test change

The change to require the DMA length to be only word-aligned was not
safe.

[XFS] remove an over-zealous WARN_ON

Merge ... /linux/kernel/git/torvalds/linux-2.6.git

libata: Fix use-after-iounmap

Jens Axboe pointed out that the iounmap() call in libata was occurring
too early, and some drivers (ahci, probably others) were using ioremap'd
memory after it had been unmapped.

The patch should address that problem by way of improving the libata
driver API:

* move ->host_stop() call after all ->port_stop() calls have occurred.

* create default helper function ata_host_stop(), and move iounmap()
call there.

* add ->host_stop_prewalk() hook, use it in sata_qstor.c (hi Mark).
sata_qstor appears to require the host-stop-before-port-stop ordering
that existed prior to applying the attached patch.

[PATCH] i386: fix prevent_tail_call

We fixed this bug before, but it didn't take.  It may have been the case
that the problem was first noticed to occur in a CONFIG_REGPARM compile.
But it's not regparm functions that need not to make tail calls, it's
asmlinkage functions called with a user pt_regs frame on the stack
supplying their arguments.  prevent_tail_call probably doesn't do anything
at all in regparm functions (your argument registers are going to be
clobbered, period).  It was a braino to conditionalize that definition in
the first place.

Signed-off-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Merge ... /pub/scm/linux/kernel/git/davem/net-2.6