Pull drm fixes from Dave Airlie:
 "As soon as I sent the non-urgent stack, two important fixes come in:

   - i915: fixes SNB GPU hangs in a number of 3D apps

   - radeon: initial fix for VGA on LLano system, 3 or 4 of us have
     spent time debugging this, and Jerome finally figured out the magic
     bit the BIOS/fglrx set that we didn't.  This at least should get
     things working, there may be future reliability fixes."

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  drm/i915: Set the Stencil Cache eviction policy to non-LRA mode.
  drm/radeon/kms: need to set up ss on DP bridges as well

This reverts commit a32744d4.

While that commit was technically the right thing to do, and made the
x86-64 compat mode work identically to native 32-bit mode (and thus
fixing the problem with a 32-bit systemd install on a 64-bit kernel), it
turns out that the automount binaries had workarounds for this compat
problem.

Now, the workarounds are disgusting: doing an "uname()" to find out the
architecture of the kernel, and then comparing it for the 64-bit cases
and fixing up the size of the read() in automount for those.  And they
were confused: it's not actually a generic 64-bit issue at all, it's
very much tied to just x86-64, which has different alignment for an
'u64' in 64-bit mode than in 32-bit mode.

But the end result is that fixing the compat layer actually breaks the
case of a 32-bit automount on a x86-64 kernel.

There are various approaches to fix this (including just doing a
"strcmp()" on current->comm and comparing it to "automount"), but I
think that I will do the one that teaches pipes about a special "packet
mode", which will allow user space to not have to care too deeply about
the padding at the end of the autofs packet.

That change will make the compat workaround unnecessary, so let's revert
it first, and get automount working again in compat mode.  The
packetized pipes will then fix autofs for systemd.
Reported-and-requested-by: Michael Tokarev <mjt@tls.msk.ru>
Cc: Ian Kent <raven@themaw.net>
Cc: stable@kernel.org # for 3.3
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Clearing bit 5 of CACHE_MODE_0 is necessary to prevent GPU hangs in
OpenGL programs such as Google MapsGL, Google Earth, and gzdoom when
using separate stencil buffers.  Without it, the GPU tries to use the
LRA eviction policy, which isn't supported.  This was supposed to be off
by default, but seems to be on for many machines.

This cannot be done in gen6_init_clock_gating with most of the other
workaround bits; the render ring needs to exist.  Otherwise, the
register write gets dropped on the floor (one printk will show it
changed, but a second printk immediately following shows the value
reverts to the old one).

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47535
Cc: stable@vger.kernel.org
Cc: Rob Castle <futuredub@gmail.com>
Cc: Eric Appleman <erappleman@gmail.com>
Cc: aaron667@gmx.net
Cc: Keith Packard <keithp@keithp.com>
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>

Makes Nutmeg DP to VGA bridges work for me.

Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=42490

Noticed by Jerome Glisse (after weeks of debugging).
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>

Pull CIFS fixes from Steve French.

* git://git.samba.org/sfrench/cifs-2.6:
  Use correct conversion specifiers in cifs_show_options
  CIFS: Show backupuid/gid in /proc/mounts
  cifs: fix offset handling in cifs_iovec_write

Merge tag 'stable/for-linus-3.4-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen

Pull Xen fixes from Konrad Rzeszutek Wilk:
 "Some of these had been in existence since the 2.6.27 days, some since
  3.0 - and some due to new features added in v3.4.

  The one that is most interesting is David's one - in the low-level
  assembler code we had be checking events needlessly.  With his patch
  now we do it when the appropriate flag is set - with the added benefit
  that we can process events faster.  Stefano's is fixing a mistake
  where the Linux IRQ numbers were ACK-ed instead of the Xen IRQ,
  resulting in missing interrupts.  The other ones are bootup related
  that can show up on various hardware."

 - In the low-level assembler code we would jump to check events even if
   none were present.  This incorrect behavior had been there since
   2.6.27 days!
 - When using the fast-path for ACK-ing interrupts we were using the
   Linux IRQ numbers instead of the Xen ones (and they can differ) and
   missing interrupts in process.
 - Fix bootup crashes when ACPI hotplug CPUs were present and they would
   expand past the set number of CPUs we were allocated.
 - Deal with broken BIOSes when uploading C-states to the hypervisor.
 - Disable the cpuid check for MWAIT_LEAF if the ACPI PAD driver is
   loaded.  If the ACPI PAD driver is used it will crash, so lets not
   export the functionality so the ACPI PAD driver won't load.

* tag 'stable/for-linus-3.4-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen:
  xen: correctly check for pending events when restoring irq flags
  xen/acpi: Workaround broken BIOSes exporting non-existing C-states.
  xen/smp: Fix crash when booting with ACPI hotplug CPUs.
  xen: use the pirq number to check the pirq_eoi_map
  xen/enlighten: Disable MWAIT_LEAF so that acpi-pad won't be loaded.

Pull misc SPI device driver bug fixes from Grant Likely.

* tag 'spi-for-linus' of git://git.secretlab.ca/git/linux-2.6:
  spi/spi-bfin5xx: Fix flush of last bit after each spi transfer
  spi/spi-bfin5xx: fix reversed if condition in interrupt mode
  spi/spi_bfin_sport: drop bits_per_word from client data
  spi/bfin_spi: drop bits_per_word from client data
  spi/spi-bfin-sport: move word length setup to transfer handler
  spi/bfin5xx: rename config macro name for bfin5xx spi controller driver
  spi/pl022: Allow request for higher frequency than maximum possible
  spi/bcm63xx: set master driver mode_bits.
  spi/bcm63xx: don't use the stopping state
  spi/bcm63xx: convert to the pump message infrastructure
  spi/spi-ep93xx.c: use dma_transfer_direction instead of dma_data_direction
  spi: fix spi.h kernel-doc warning
  spi/pl022: Fix calculate_effective_freq()
  spi/pl022: Fix range checking for bits per word

Pull hwmon patches from Guenter Roeck:
 - Fix build warning in ad7314 driver
 - Fix pci_device_id array access in fam15h_power driver, introduced by
   commit 00250ec9 ("hwmon: fam15h_power: fix bogus values with
   current BIOSes")

* tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (fam15h_power) Fix pci_device_id array
  hwmon: (ad7314) Fix build warning

Pull drm fixes from Dave Airlie:
 "For your Friday pull request stack, nothing astounding or shattering
  this week some exynos, some intel, some radeon fixes.  One intel fix
  for a regression somwehere back in 2.6.35 land."

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  drm/radeon/kms: use frac fb div on APUs
  drm/radeon: add a missing entry to encoder_names
  drm/i915: handle input/output sdvo timings separately in mode_set
  drm/i915: fix integer overflow in i915_gem_do_execbuffer()
  drm/i915: fix integer overflow in i915_gem_execbuffer2()
  drm/exynos: added missed vm area region mapping type.
  drm/exynos: fixed exynos_drm_gem_map_pages bug.
  drm/exynos: fixed duplicatd memory allocation bug.
  drm/i915: fixup load-detect on enabled, but not active pipe

Pull RCU fix from Ingo Molnar.

* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  rcu: Permit call_rcu() from CPU_DYING notifiers

Pull x86 fixes from Ingo Molnar.

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/apic: Use x2apic physical mode based on FADT setting
  x86/mrst: Quiet sparse noise about plain integer as NULL pointer
  x86, intel_cacheinfo: Fix error return code in amd_set_l3_disable_slot()

Pull scheduler fixes from Ingo Molnar.

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched: Fix OOPS when build_sched_domains() percpu allocation fails
  sched: Fix more load-balancing fallout

Pull perf fixes from Ingo Molnar.

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf: Fix perf_event_for_each() to use sibling
  perf symbols: Read plt symbols from proper symtab_type binary
  tracing: Fix stacktrace of latency tracers (irqsoff and friends)
  perf tools: Add 'G' and 'H' modifiers to event parsing
  tracing: Fix regression with tracing_on
  perf tools: Drop CROSS_COMPILE from flex and bison calls
  perf report: Fix crash showing warning related to kernel maps
  tracing: Fix build breakage without CONFIG_PERF_EVENTS (again)

Pull build fixes for less mainstream architectures from Paul Gortmaker:
 "These are fixes for frv(1), blackfin(2), powerpc(1) and xtensa(4).

  Fortunately the touches are nearly all specific to files just used by
  the arch in question.  The two touches to shared/common files
  [kernel/irq/debug.h and drivers/pci/Makefile] are trivial to assess as
  no risk to anyone.

  Half of them relate to xtensa directly.  It was only when I fixed the
  last xtensa issue that I realized that the arch has been broken for a
  significant time, and isn't a specific v3.4 regression.  So if you
  wanted, we could leave xtensa lying bleeding in the street for a
  couple more weeks and queue those for 3.5.  But given they are no risk
  to anyone outside of xtensa, I figured to just leave them in.

  If you are OK with taking the xtensa fixes, then please pull to get:

   - one last implicit include uncovered by system.h that is in a file
     specific to just one powerpc defconfig.  (I'd sync'd with BenH).

   - fix an oversight in the PCI makefile where shared code wasn't being
     compiled for ARCH=frv

   - fix a missing include for GPIO in blackfin framebuffer.

   - audit and tag endif in blackfin ezkit board file, in order to find
     and fix the misplaced endif masking a block of code.

   - fix irq/debug.h choice of temporary macro names to be more internal
     so they don't conflict with names used by xtensa.

   - fix a reference to an undeclared local var in xtensa's signal.c

   - fix an implicit bug.h usage in xtensa's asm/io.h uncovered by my
     removing bug.h from kernel.h

   - fix xtensa to properly indicate it is using asm-generic/hardirq.h
     in order to resolve the link error - undefined ack_bad_irq

  The xtensa still fails final link as my latest binutils does something
  evil when ld forward-relocates unlikely() blocks, but in theory people
  who have older/valid toolchains could now use the thing."

* 'for-v3.4-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/paulg/linux:
  xtensa: fix build fail on undefined ack_bad_irq
  blackfin: fix ifdef fustercluck in mach-bf538/boards/ezkit.c
  blackfin: fix compile error in bfin-lq035q1-fb.c
  pci: frv architecture needs generic setup-bus infrastructure
  irq: hide debug macros so they don't collide with others.
  xtensa: fix build error in xtensa/include/asm/io.h
  xtensa: fix build failure in xtensa/kernel/signal.c
  powerpc: fix system.h fallout in sysdev/scom.c [chroma_defconfig]

Pull SuperH fixes from Paul Mundt.

* tag 'sh-for-linus' of git://github.com/pmundt/linux-sh:
  sh: Fix up tracepoint build fallout from static key introduction.
  sh: Fix error synchronising kernel page tables

Pull security key doc update from Jeff Layton:
 "Ordinarily, I send my patches through others' trees, but David
  suggested I just send this one to you directly since it's just a
  Documentation/ update"

* 'docs-3.4' of git://git.samba.org/jlayton/linux:
  keys: update the documentation with info about "logon" keys

In xen_restore_fl_direct(), xen_force_evtchn_callback() was being
called even if no events were pending.  This resulted in (depending on
workload) about a 100 times as many xen_version hypercalls as
necessary.

Fix this by correcting the sense of the conditional jump.

This seems to give a significant performance benefit for some
workloads.

There is some subtle tricksy "..since the check here is trying to
check both pending and masked in a single cmpw, but I think this is
correct. It will call check_events now only when the combined
mask+pending word is 0x0001 (aka unmasked, pending)." (Ian)

CC: stable@kernel.org
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>

This patch ensures that the last bit of a transfer gets correctly
flushed out of the register.
Signed-off-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

This condition is used to determine 8 bits or 16 and 32 bits transfer.
Obviously it is reversed.
Signed-off-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

Since the member was dropped from the common Blackfin header, we need
to stop using it in the SPORT driver too.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

No other SPI controller has this field, and SPI clients should be setting
this up in their own drivers.  So drop it from the Blackfin controller to
keep people from using it.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

Each transfer may have its own bits per word.
Signed-off-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

This controller is only for blackfin 5xx soc, so rename it to BFIN5XX
Signed-off-by: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

Currently, if we request for frequency greater than maximum possible, spi driver
returns error.

For example, if the spi block src frequency is 333/4 MHz, i.e. 83.33.. MHz,
maximum frequency programmable would be src/2. Which would come around 41.6...

It is difficult to pass frequency in these figures. We normally try to program
in round figures, like 42 MHz and it should get programmed to <=
requested_frequency, i.e. 41.6...

For this to happen, we must not return error even if requested freq is higher
than max possible. But should program it to max possible.
Reported-by: Vinit Kamalaksha Shenoy <vinit.shenoy@st.com>
Signed-off-by: Viresh Kumar <viresh.kumar@st.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

We were not properly advertising the MODE bits supported by this driver, fix
that.
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

We do not need to use a flag to indicate if the master driver is stopping
it is sufficient to perform spi master unregistering in the platform
driver's remove function.
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

This patch converts the bcm63xx SPI driver to use the SPI infrastructure
pump message queue. Since we were previously sleeping in the SPI
driver's transfer() function (which is not allowed) this is now fixed as well.

To complete that conversion a certain number of changes have been made:
- the transfer len is split into multiple hardware transfers in case its
  size is bigger than the hardware FIFO size
- the FIFO refill is no longer done in the interrupt context, which was a
  bad idea leading to quick interrupt handler re-entrancy
Tested-by: Tanguy Bouzeloc <tanguy.bouzeloc@efixo.com>
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

A new enum indicating the dma channel direction was introduced by:

commit 49920bc6
    dmaengine: add new enum dma_transfer_direction

The following commit changed spi-ep93xx to use the new enum:

commit a485df4b
    spi, serial: move to dma_transfer_direction

In doing so a sparse warning was introduced:

warning: mixing different enum types
   int enum dma_data_direction  versus
   int enum dma_transfer_direction

This is produced because the 'dir' passed in ep93xx_spi_dma_prepare
is an enum dma_data_direction and is being used to set the
dma_slave_config 'direction' which is now an enum dma_transfer_direction.

Fix this by converting spi-ep93xx to use the new enum type in all
places.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Acked-by: Mika Westerberg <mika.westerberg@iki.fi>
Acked-by: Vinod Koul <vinod.koul@intel.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

Fix kernel-doc warning in spi.h (copy/paste):

Warning(include/linux/spi/spi.h:365): No description found for parameter 'unprepare_transfer_hardware'
Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

calculate_effective_freq() was still not optimized and there were cases when it
returned without error and with values of cpsr and scr as zero.

Also, the variable named found is not used well.

This patch targets to optimize and correct this routine. Tested for SPEAr.
Signed-off-by: Viresh Kumar <viresh.kumar@st.com>
Tested-by: Vinit Kamalaksha Shenoy <vinit.shenoy@st.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

Seems to be more stable on certain monitors.

Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=48880Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>

An entry for INTERNAL_VCE encoder was missing. Add it.
Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

From Daniel Vetter

- VGA load-detect fix. This bug seems to be as old as the load-detect code
  (2.6.30), but needs stupid userspace (upowerd trying to detect
  connectors on dpms-off outputs) to actually kill the machine. And
  obviously a machine without VGA-hotplug, otherwise we don't do load
  detect.
- 2 interger overflow fixes for unpriviledged ioctls from Xi Wang.
- Fix SDVO regression for low-res (pixelclock < 100MHz) digital outputs,
 introduce in 2.6.36.

* 'drm-intel-fixes' of git://people.freedesktop.org/~danvet/drm-intel:
  drm/i915: handle input/output sdvo timings separately in mode_set
  drm/i915: fix integer overflow in i915_gem_do_execbuffer()
  drm/i915: fix integer overflow in i915_gem_execbuffer2()
  drm/i915: fixup load-detect on enabled, but not active pipe

Merge branch 'exynos-drm-fixes' of git://git.infradead.org/users/kmpark/linux-2.6-samsung into drm-fixes

From Inki Dae:

this patch set fixes gem allocation and mapping issue between user space and
physical memory region.

* 'exynos-drm-fixes' of git://git.infradead.org/users/kmpark/linux-2.6-samsung:
  drm/exynos: added missed vm area region mapping type.
  drm/exynos: fixed exynos_drm_gem_map_pages bug.
  drm/exynos: fixed duplicatd memory allocation bug.

We did a similar check for the P-states but did not do it for
the C-states. What we want to do is ignore cases where the DSDT
has definition for sixteen CPUs, but the machine only has eight
CPUs and we get:
xen-acpi-processor: (CX): Hypervisor error (-22) for ACPI CPU14
Reported-by: Tobias Geiger <tobias.geiger@vido.info>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>

When we boot on a machine that can hotplug CPUs and we
are using 'dom0_max_vcpus=X' on the Xen hypervisor line
to clip the amount of CPUs available to the initial domain,
we get this:

(XEN) Command line: com1=115200,8n1 dom0_mem=8G noreboot dom0_max_vcpus=8 sync_console mce_verbosity=verbose console=com1,vga loglvl=all guest_loglvl=all
.. snip..
DMI: Intel Corporation S2600CP/S2600CP, BIOS SE5C600.86B.99.99.x032.072520111118 07/25/2011
.. snip.
SMP: Allowing 64 CPUs, 32 hotplug CPUs
installing Xen timer for CPU 7
cpu 7 spinlock event irq 361
NMI watchdog: disabled (cpu7): hardware events not enabled
Brought up 8 CPUs
.. snip..
	[acpi processor finds the CPUs are not initialized and starts calling
	arch_register_cpu, which creates /sys/devices/system/cpu/cpu8/online]
CPU 8 got hotplugged
CPU 9 got hotplugged
CPU 10 got hotplugged
.. snip..
initcall 1_acpi_battery_init_async+0x0/0x1b returned 0 after 406 usecs
calling  erst_init+0x0/0x2bb @ 1

	[and the scheduler sticks newly started tasks on the new CPUs, but
	said CPUs cannot be initialized b/c the hypervisor has limited the
	amount of vCPUS to 8 - as per the dom0_max_vcpus=8 flag.
	The spinlock tries to kick the other CPU, but the structure for that
	is not initialized and we crash.]
BUG: unable to handle kernel paging request at fffffffffffffed8
IP: [<ffffffff81035289>] xen_spin_lock+0x29/0x60
PGD 180d067 PUD 180e067 PMD 0
Oops: 0002 [#1] SMP
CPU 7
Modules linked in:

Pid: 1, comm: swapper/0 Not tainted 3.4.0-rc2upstream-00001-gf5154e8 #1 Intel Corporation S2600CP/S2600CP
RIP: e030:[<ffffffff81035289>]  [<ffffffff81035289>] xen_spin_lock+0x29/0x60
RSP: e02b:ffff8801fb9b3a70  EFLAGS: 00010282

With this patch, we cap the amount of vCPUS that the initial domain
can run, to exactly what dom0_max_vcpus=X has specified.

In the future, if there is a hypercall that will allow a running
domain to expand past its initial set of vCPUS, this patch should
be re-evaluated.

CC: stable@kernel.org
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>

With the introduction of static keys, anything using tracepoints blows up
in the following manner:

include/trace/events/oom.h:8:13: error: initializer element is not constant
include/trace/events/oom.h:8:13: error: (near initialization for '__tracepoint_oom_score_adj_update')
include/trace/events/oom.h:8:13: error: initializer element is not constant
include/trace/events/oom.h:8:13: error: (near initialization for '__tracepoint_oom_score_adj_update.key')

This is a result of the STATIC_KEY_INIT_xxx defs wrapping ATOMIC_INIT()
which on sh includes an atomic_t typecast. Given that we don't really
need the typecast for anything anymore, the simplest solution is simply
to kill off the cast.
Signed-off-by: Nobuhiro Iwamatsu <nobuhiro.iwamatsu.yj@renesas.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

Pull arch/tile fixes from Chris Metcalf:
 "One change fixes a platform-independent bug about environment var
  handling in the boot command line.  The other is a trivial
  tile-specific bug fix to avoid a link-time warning."

* 'stable' of git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile:
  arch/tile: fix a couple of functions that should be __init
  init: fix bug where environment vars can't be passed via boot args

Pull infiniband fixes from Roland Dreier:
 "A few fixes for regressions introduced in 3.4-rc1:
   - fix memory leak in mlx4
   - fix two problems with new MAD response generation code"

* tag 'ib-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  IB/mlx4: Fix memory leaks in ib_link_query_port()
  IB/mad: Don't send response for failed MADs
  IB/mad: Set 'D' bit in response for unhandled MADs