...this string is zeroed out and nothing ever changes it.
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NSteve French <sfrench@us.ibm.com>

Some of the code under CONFIG_CIFS_ACL is dependent upon code under
CONFIG_CIFS_EXPERIMENTAL, but the Kconfig options don't reflect that
dependency. Move more of the ACL code out from under
CONFIG_CIFS_EXPERIMENTAL and under CONFIG_CIFS_ACL.

Also move find_readable_file out from other any sort of Kconfig
option and make it a function normally compiled in.
Reported-and-Acked-by: NRandy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NSteve French <sfrench@us.ibm.com>

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
  HID: length resolution should be reported units/mm
  HID: add support for F430 Force Feedback Wheel
  HID: egalax: Use kzalloc
  HID: Remove KERN_DEBUG from dbg_hid use

Manually fixed trivial conflict in drivers/hid/hid-input.c (due to
removal of KERN_DEBUG from dbg_hid use clashing with new keycode
interface switch)

If a user manages to trigger an oops with fs set to KERNEL_DS, fs is not
otherwise reset before do_exit().  do_exit may later (via mm_release in
fork.c) do a put_user to a user-controlled address, potentially allowing
a user to leverage an oops into a controlled write into kernel memory.

This is only triggerable in the presence of another bug, but this
potentially turns a lot of DoS bugs into privilege escalations, so it's
worth fixing.  I have proof-of-concept code which uses this bug along
with CVE-2010-3849 to write a zero to an arbitrary kernel address, so
I've tested that this is not theoretical.

A more logical place to put this fix might be when we know an oops has
occurred, before we call do_exit(), but that would involve changing
every architecture, in multiple places.

Let's just stick it in do_exit instead.

[akpm@linux-foundation.org: update code comment]
Signed-off-by: NNelson Elhage <nelhage@ksplice.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: <stable@kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

commit 62b61f61 ("ksm: memory hotremove migration only") caused the
following new lockdep warning.

  =======================================================
  [ INFO: possible circular locking dependency detected ]
  -------------------------------------------------------
  bash/1621 is trying to acquire lock:
   ((memory_chain).rwsem){.+.+.+}, at: [<ffffffff81079339>]
  __blocking_notifier_call_chain+0x69/0xc0

  but task is already holding lock:
   (ksm_thread_mutex){+.+.+.}, at: [<ffffffff8113a3aa>]
  ksm_memory_callback+0x3a/0xc0

  which lock already depends on the new lock.

  the existing dependency chain (in reverse order) is:

  -> #1 (ksm_thread_mutex){+.+.+.}:
       [<ffffffff8108b70a>] lock_acquire+0xaa/0x140
       [<ffffffff81505d74>] __mutex_lock_common+0x44/0x3f0
       [<ffffffff81506228>] mutex_lock_nested+0x48/0x60
       [<ffffffff8113a3aa>] ksm_memory_callback+0x3a/0xc0
       [<ffffffff8150c21c>] notifier_call_chain+0x8c/0xe0
       [<ffffffff8107934e>] __blocking_notifier_call_chain+0x7e/0xc0
       [<ffffffff810793a6>] blocking_notifier_call_chain+0x16/0x20
       [<ffffffff813afbfb>] memory_notify+0x1b/0x20
       [<ffffffff81141b7c>] remove_memory+0x1cc/0x5f0
       [<ffffffff813af53d>] memory_block_change_state+0xfd/0x1a0
       [<ffffffff813afd62>] store_mem_state+0xe2/0xf0
       [<ffffffff813a0bb0>] sysdev_store+0x20/0x30
       [<ffffffff811bc116>] sysfs_write_file+0xe6/0x170
       [<ffffffff8114f398>] vfs_write+0xc8/0x190
       [<ffffffff8114fc14>] sys_write+0x54/0x90
       [<ffffffff810028b2>] system_call_fastpath+0x16/0x1b

  -> #0 ((memory_chain).rwsem){.+.+.+}:
       [<ffffffff8108b5ba>] __lock_acquire+0x155a/0x1600
       [<ffffffff8108b70a>] lock_acquire+0xaa/0x140
       [<ffffffff81506601>] down_read+0x51/0xa0
       [<ffffffff81079339>] __blocking_notifier_call_chain+0x69/0xc0
       [<ffffffff810793a6>] blocking_notifier_call_chain+0x16/0x20
       [<ffffffff813afbfb>] memory_notify+0x1b/0x20
       [<ffffffff81141f1e>] remove_memory+0x56e/0x5f0
       [<ffffffff813af53d>] memory_block_change_state+0xfd/0x1a0
       [<ffffffff813afd62>] store_mem_state+0xe2/0xf0
       [<ffffffff813a0bb0>] sysdev_store+0x20/0x30
       [<ffffffff811bc116>] sysfs_write_file+0xe6/0x170
       [<ffffffff8114f398>] vfs_write+0xc8/0x190
       [<ffffffff8114fc14>] sys_write+0x54/0x90
       [<ffffffff810028b2>] system_call_fastpath+0x16/0x1b

But it's a false positive.  Both memory_chain.rwsem and ksm_thread_mutex
have an outer lock (mem_hotplug_mutex).  So they cannot deadlock.

Thus, This patch annotate ksm_thread_mutex is not deadlock source.

[akpm@linux-foundation.org: update comment, from Hugh]
Signed-off-by: NKOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: NHugh Dickins <hughd@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Presently hwpoison is using lock_system_sleep() to prevent a race with
memory hotplug.  However lock_system_sleep() is a no-op if
CONFIG_HIBERNATION=n.  Therefore we need a new lock.
Signed-off-by: NKOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Suggested-by: NHugh Dickins <hughd@google.com>
Acked-by: NHugh Dickins <hughd@google.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

->releasepage() does not remove the page from the mapping.
Acked-by: NNeil Brown <neilb@suse.de>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

On stock 2.6.37-rc4, running:

  # mount lilith:/export /mnt/lilith
  # find  /mnt/lilith/ -type f -print0 | xargs -0 file

crashes the machine fairly quickly under Xen.  Often it results in oops
messages, but the couple of times I tried just now, it just hung quietly
and made Xen print some rude messages:

    (XEN) mm.c:2389:d80 Bad type (saw 7400000000000001 != exp
    3000000000000000) for mfn 1d7058 (pfn 18fa7)
    (XEN) mm.c:964:d80 Attempt to create linear p.t. with write perms
    (XEN) mm.c:2389:d80 Bad type (saw 7400000000000010 != exp
    1000000000000000) for mfn 1d2e04 (pfn 1d1fb)
    (XEN) mm.c:2965:d80 Error while pinning mfn 1d2e04

Which means the domain tried to map a pagetable page RW, which would
allow it to map arbitrary memory, so Xen stopped it.  This is because
vm_unmap_ram() left some pages mapped in the vmalloc area after NFS had
finished with them, and those pages got recycled as pagetable pages
while still having these RW aliases.

Removing those mappings immediately removes the Xen-visible aliases, and
so it has no problem with those pages being reused as pagetable pages.
Deferring the TLB flush doesn't upset Xen because it can flush the TLB
itself as needed to maintain its invariants.

When unmapping a region in the vmalloc space, clear the ptes
immediately.  There's no point in deferring this because there's no
amortization benefit.

The TLBs are left dirty, and they are flushed lazily to amortize the
cost of the IPIs.

This specific motivation for this patch is an oops-causing regression
since 2.6.36 when using NFS under Xen, triggered by the NFS client's use
of vm_map_ram() introduced in 56e4ebf8 ("NFS: readdir with vmapped
pages") .  XFS also uses vm_map_ram() and could cause similar problems.
Signed-off-by: NJeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Nick Piggin <npiggin@kernel.dk>
Cc: Bryan Schumaker <bjschuma@netapp.com>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: Alex Elder <aelder@sgi.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The AMD Geode CS5536 Companion Device Silicon Revision B1 Specification
Update mentions the follow as issue #36:

 "Atomic write transactions to the atomic GPIO High Bank Feature Bit
  registers should only affect the bits selected [...]"

 "after Suspend, an atomic write transaction [...] will clear all
  non-selected bits of the accessed register."

In other words, writing to the high bank for a single GPIO bit will
clear every other GPIO bit (but only sometimes after a suspend).

The workaround described is obvious and simple; do a read-modify-write.
This patch does that, and documents why we're doing it.
Signed-off-by: NAndres Salomon <dilinger@queued.net>
Cc: <stable@kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

reiserfs_acl_chmod() can be called by reiserfs_set_attr() and then take
the reiserfs lock a second time.  Thereafter it may call journal_begin()
that definitely requires the lock not to be nested in order to release
it before taking the journal mutex because the reiserfs lock depends on
the journal mutex already.

So, aviod nesting the lock in reiserfs_acl_chmod().
Reported-by: NPawel Zawora <pzawora@gmail.com>
Signed-off-by: NFrederic Weisbecker <fweisbec@gmail.com>
Tested-by: NPawel Zawora <pzawora@gmail.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Cc: <stable@kernel.org>		[2.6.32.x+]
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

It's not useful to build LED triggers when there's no LEDs that can be
triggered by them.  Therefore, fix up the dependencies so that this
cannot happen, and fix a few users that select triggers to depend on
LEDS_CLASS as well (there is also one user that also selects LEDS_CLASS,
which is OK).
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Reported-by: NRandy Dunlap <randy.dunlap@oracle.com>
Acked-by: NRandy Dunlap <randy.dunlap@oracle.com>
Tested-by: NIngo Molnar <mingo@elte.hu>
Cc: Arnd Hannemann <arnd@arndnet.de>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Richard Purdie <rpurdie@linux.intel.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The nr_dirty_[background_]threshold fields are misplaced before the
numa_* fields, and users will read strange values.

This is the right order.  Before patch, nr_dirty_background_threshold
will read as 0 (the value from numa_miss).

	numa_hit 128501
	numa_miss 0
	numa_foreign 0
	numa_interleave 7388
	numa_local 128501
	numa_other 0
	nr_dirty_threshold 144291
	nr_dirty_background_threshold 72145
Signed-off-by: NWu Fengguang <fengguang.wu@intel.com>
Cc: Michael Rubin <mrubin@google.com>
Reviewed-by: NKOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Reviewed-by: NMinchan Kim <minchan.kim@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

find_task_by_vpid() should be protected by rcu_read_lock(), to prevent
free_pid() reclaiming pid.
Signed-off-by: NZeng Zhaoming <zengzm.kernel@gmail.com>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Have hugetlb_fault() call unlock_page(page) only if it had previously
called lock_page(page).

Setting CONFIG_DEBUG_VM=y and then running the libhugetlbfs test suite,
resulted in the tripping of VM_BUG_ON(!PageLocked(page)) in
unlock_page() having been called by hugetlb_fault() when page ==
pagecache_page.  This patch remedied the problem.
Signed-off-by: NDean Nelson <dnelson@redhat.com>
Cc: <stable@kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

* 'staging-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging-2.6: (27 commits)
  Staging: rt2870: Add USB ID for Buffalo Airstation WLI-UC-GN
  staging: easycap needs smp_lock.h, fixes build error
  Staging: batman-adv: ensure that eth_type_trans gets linear memory
  Staging: batman-adv: Don't remove interface with spinlock held
  staging: brcm80211: updated maintainers contact information
  staging: fix winbond build, needs delay.h
  Staging: line6: fix up my fixup for some sysfs attribute permissions
  Staging: zram: fix up my fixup for some sysfs attribute permissions
  Staging: udlfb: fix up my fixup for some sysfs attribute permissions
  Staging: samsung-laptop: fix up my fixup for some sysfs attribute permissions
  Staging: iio: adis16220: fix up my fixup for some sysfs attribute permissions
  Staging: frontier: fix up my fixup for some sysfs attribute permissions
  Staging: asus_oled: fix up my fixup for some sysfs attribute permissions
  staging: spectra: fix build error
  Staging: intel_sst: fix memory leak
  Staging: rtl8712: signedness bug in init
  staging: rtl8187se: Change panic to warn when RF switch turned off
  staging: comedi: fix memory leak
  Staging: quickstart: free after input_unregister_device()
  Staging: speakup: free after input_unregister_device()
  ...

Merge branch 'driver-core-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core-2.6

* 'driver-core-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core-2.6:
  uio: Change mail address of Hans J. Koch
  driver core: prune docs about device_interface
  driver core: the development tree has switched to git

* 'tty-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty-2.6:
  serial: mfd: adjust the baud rate setting
  TTY: open/hangup race fixup
  TTY: don't allow reopen when ldisc is changing
  NET: wan/x25, fix ldisc->open retval
  TTY: ldisc, fix open flag handling
  serial8250: Mark console as CON_ANYTIME

* 'usb-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6:
  USB: fix autosuspend bug in usb-serial
  USB: ehci: disable LPM and PPCD for nVidia MCP89 chips
  USB: serial: ftdi_sio: Vardaan USB RS422/485 converter PID added
  USB: yurex: add .llseek fop to file_operations
  USB: ftdi_sio: Add ID for RT Systems USB-29B radio cable
  usb: musb: do not use dma for control transfers
  usb: musb: gadget: fix compilation warning
  usb: musb: clear RXCSR_AUTOCLEAR before PIO read
  usb: musb: unmap dma buffer when switching to PIO
  xhci: Don't let the USB core disable SuperSpeed ports.
  xhci: Setup array of USB 2.0 and USB 3.0 ports.
  xhci: Fix reset-device and configure-endpoint commands

* git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog:
  watchdog: it8712f_wdt: add note to Kconfig
  watchdog: gef_wdt: include fs.h
  watchdog: bcm63xx_wdt: improve platform part.
  watchdog: iTCO_wdt: TCO Watchdog patch for Intel Patsburg PCH

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  IB: Fix information leak in marshalling code
  IB/pack: Remove some unused code added by the IBoE patches
  IB/mlx4: Fix IBoE link state
  IB/mlx4: Fix IBoE reported link rate
  mlx4_core: Workaround firmware bug in query dev cap
  IB/mlx4: Fix memory ordering of VLAN insertion control bits
  MAINTAINERS: Update NetEffect entry

Currently, the attribute cache timeout for CIFS is hardcoded to 1 second. This
means that the client might have to issue a QPATHINFO/QFILEINFO call every 1
second to verify if something has changes, which seems too expensive. On the
other hand, if the timeout is hardcoded to a higher value, workloads that
expect strict cache coherency might see unexpected results.

Making attribute cache timeout as a tunable will allow us to make a tradeoff
between performance and cache metadata correctness depending on the
application/workload needs.

Add 'actimeo' tunable that can be used to tune the attribute cache timeout.
The default timeout is set to 1 second. Also, display actimeo option value in
/proc/mounts.

It appears to me that 'actimeo' and the proposed (but not yet merged)
'strictcache' option cannot coexist, so care must be taken that we reset the
other option if one of them is set.

Changes since last post:
   - fix option parsing and handle possible values correcly
Reviewed-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NSuresh Jayaraman <sjayaraman@suse.de>
Signed-off-by: NSteve French <sfrench@us.ibm.com>

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  xfs: only run xfs_error_test if error injection is active
  xfs: avoid moving stale inodes in the AIL
  xfs: delayed alloc blocks beyond EOF are valid after writeback
  xfs: push stale, pinned buffers on trylock failures
  xfs: fix failed write truncation handling.

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lrg/voltage-2.6:
  regulator: fix kernel-doc for set_consumer_device_supply
  regulator: enable supply regulator only when use count is zero
  regulator: twl-regulator - fix twlreg_set_mode
  regulator: lock supply in regulator enable
  regulator: Return proper error for regulator_register()
  regulator: Ensure enough delay time for enabling regulator
  regulator: Remove a redundant device_remove_file call in create_regulator
  regulator: Staticise mc13783_powermisc_rmw()
  regulator: regulator disable supply fix

* 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-2.6:
  [media] v4l: Remove module_name argument to the v4l2_i2c_new_subdev* functions
  [media] v4l: Remove hardcoded module names passed to v4l2_i2c_new_subdev* (2)

* 'rbd-sysfs' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
  rbd: replace the rbd sysfs interface

* git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6:
  cifs: fix parsing of hostname in dfs referrals
  cifs: display fsc in /proc/mounts
  cifs: enable fscache iff fsc mount option is used explicitly
  cifs: allow fsc mount option only if CONFIG_CIFS_FSCACHE is set
  cifs: Handle extended attribute name cifs_acl to generate cifs acl blob (try #4)
  cifs: Misc. cleanup in cifsacl handling [try #4]
  cifs: trivial comment fix for cifs_invalidate_mapping
  [CIFS] fs/cifs/Kconfig: CIFS depends on CRYPTO_HMAC
  cifs: don't take extra tlink reference in initiate_cifs_search
  cifs: Percolate error up to the caller during get/set acls [try #4]
  cifs: fix another memleak, in cifs_root_iget
  cifs: fix potential use-after-free in cifs_oplock_break_put

On some motherboards the it8712f watchdog does not work unless
the game port was enabled. see Bug 13140. We therefor add a note
to Kconfig.
Signed-off-by: NWim Van Sebroeck <wim@iguana.be>

Add missing include "linux/fs.h".
This fixes compile failure.
Signed-off-by: NWolfram Sang <w.sang@pengutronix.de>
Signed-off-by: NMartyn Welch <martyn.welch@ge.com>
Signed-off-by: NWim Van Sebroeck <wim@iguana.be>

* fix devinit and devexit sections
* fix platform removal code so that the iounmap happens after the removal of the timer.
* changes the reboot_notifier by a platform shutdown method.
Signed-off-by: NWim Van Sebroeck <wim@iguana.be>

This patch adds an additional LPC Controller DeviceID for the Intel Patsburg PCH for TCO Watchdog.
Signed-off-by: NSeth Heasley <seth.heasley@intel.com>
Signed-off-by: NWim Van Sebroeck <wim@iguana.be>

Input ABI requires reporting resolution on main axes in units per
millimeter, not units per inch, so we need to convert accordingly.
Tested-by: NNikolai Kondrashov <spbnick@gmail.com>
Acked-by: NNikolai Kondrashov <spbnick@gmail.com>
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

ib_ucm_init_qp_attr() and ucma_init_qp_attr() pass struct ib_uverbs_qp_attr
with reserved, qp_state, {ah_attr,alt_ah_attr}{reserved,->grh.reserved}
fields uninitialized to copy_to_user().  This leads to leaking of
contents of kernel stack memory to userspace.
Signed-off-by: NVasiliy Kulikov <segoon@openwall.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Remove unused functions added by commit ff7f5aab ("IB/pack: IBoE UD
packet packing support").
Signed-off-by: NOr Gerlitz <ogerlitz@voltaire.com>

Use netif_running() and netif_carrier_ok() to report link state,
exactly as is done to report Ethernet link state in sysfs.
Signed-off-by: NEli Cohen <eli@mellanox.co.il>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

The link rate is the product of the link speed in the link width. For
Etherent ports the rate is 10G, so we use 1 for the width and 4 for
speed to get the correct rate.
Signed-off-by: NEli Cohen <eli@mellanox.co.il>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

ConnectX firmware is supposed to report the number blue flame
registers per page as log2 of the value.  However, due to a firmware
bug, it reports actual number.  This patch works around this by
checking if the number of registers calculated fits within a page.  If
it does not, we use 8 registers per page.
Signed-off-by: NEli Cohen <eli@mellanox.co.il>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

The new interface creates directories per mapped image
and under each it creates a subdir per available snapshot.
This allows keeping a cleaner interface within the sysfs
guidelines. The ABI documentation was updated too.
Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NYehuda Sadeh <yehuda@hq.newdream.net>
Signed-off-by: NSage Weil <sage@newdream.net>

We must fully update the control segment before marking it as valid,
so that hardware doesn't start executing it before we're ready.
Signed-off-by: NEli Cohen <eli@mellanox.co.il>

[ Move VLAN control bit setting to before wmb().  - Roland ]
Signed-off-by: NRoland Dreier <rolandd@cisco.com>