linux-stable.git/drivers/net/macvlan.c, branch linux-3.13.y

macvlan: Add support for 'always_on' offload features

2014-03-24T04:44:01+00:00

[ Upstream commit 8b4703e9bd1172a5f8244276ebb94302e6153e26 ]

Macvlan currently inherits all of its features from the lower
device. When lower device disables offload support, this causes
macvlan to disable offload support as well. This causes
performance regression when using macvlan/macvtap in bridge
mode.

It can be easily demonstrated by creating 2 namespaces using
macvlan in bridge mode and running netperf between them:

MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.0.0.1 () port 0 AF_INET
Recv Send Send
Socket Socket Message Elapsed
Size Size Size Time Throughput
bytes bytes bytes secs. 10^6bits/sec

87380 16384 16384 20.00 1204.61

To restore the performance, we add software offload features
to the list of "always_on" features for macvlan. This way
when a namespace or a guest using macvtap initially sends a
packet, this packet will not be segmented at macvlan level.
It will only be segmented when macvlan sends the packet
to the lower device.

87380 16384 16384 20.00 5507.35

Fixes: 6acf54f1cf0a6747bac9fea26f34cfc5a9029523 (macvtap: Add support of packet capture on macvtap device.)
Fixes: 797f87f83b60685ff8a13fa0572d2f10393c50d3 (macvlan: fix netdev feature propagation from lower device)
CC: Florian Westphal
CC: Christian Borntraeger
CC: Jason Wang
CC: Michael S. Tsirkin
Tested-by: Christian Borntraeger
Signed-off-by: Vlad Yasevich
Signed-off-by: David S. Miller
Signed-off-by: Greg Kroah-Hartman

net: core: explicitly select a txq before doing l2 forwarding

2014-01-10T18:23:08+00:00

Currently, the tx queue were selected implicitly in ndo_dfwd_start_xmit(). The
will cause several issues:

- NETIF_F_LLTX were removed for macvlan, so txq lock were done for macvlan
  instead of lower device which misses the necessary txq synchronization for
  lower device such as txq stopping or frozen required by dev watchdog or
  control path.
- dev_hard_start_xmit() was called with NULL txq which bypasses the net device
  watchdog.
- dev_hard_start_xmit() does not check txq everywhere which will lead a crash
  when tso is disabled for lower device.

Fix this by explicitly introducing a new param for .ndo_select_queue() for just
selecting queues in the case of l2 forwarding offload. netdev_pick_tx() was also
extended to accept this parameter and dev_queue_xmit_accel() was used to do l2
forwarding transmission.

With this fixes, NETIF_F_LLTX could be preserved for macvlan and there's no need
to check txq against NULL in dev_hard_start_xmit(). Also there's no need to keep
a dedicated ndo_dfwd_start_xmit() and we can just reuse the code of
dev_queue_xmit() to do the transmission.

In the future, it was also required for macvtap l2 forwarding support since it
provides a necessary synchronization method.

Cc: John Fastabend 
Cc: Neil Horman 
Cc: e1000-devel@lists.sourceforge.net
Signed-off-by: Jason Wang 
Acked-by: Neil Horman 
Acked-by: John Fastabend 
Signed-off-by: David S. Miller

macvlan: forbid L2 fowarding offload for macvtap

2014-01-10T18:23:08+00:00

L2 fowarding offload will bypass the rx handler of real device. This will make
the packet could not be forwarded to macvtap device. Another problem is the
dev_hard_start_xmit() called for macvtap does not have any synchronization.

Fix this by forbidding L2 forwarding for macvtap.

Cc: John Fastabend 
Cc: Neil Horman 
Acked-by: Neil Horman 
Signed-off-by: Jason Wang 
Acked-by: John Fastabend 
Signed-off-by: David S. Miller

macvlan: fix netdev feature propagation from lower device

2013-12-26T18:41:00+00:00

There are inconsistencies wrt. feature propagation/inheritance between
macvlan and the underlying interface.

When a feature is turned off on the real device before a macvlan is
created on top, these will remain enabled on the macvlan device, whereas
turning off the feature on the lower device after macvlan creation the
kernel will propagate the changes to the macvlan.

The second issue is that, when propagating changes from underlying device
to the macvlan interface, macvlan can erronously lose its NETIF_F_LLTX flag,
as features are anded with the underlying device.

However, LLTX should be kept since it has no dependencies on physical
hardware (LLTX is set on macvlan creation regardless of the lower
device properties, see 8ffab51b3dfc54876f145f15b351c41f3f703195
(macvlan: lockless tx path).

The LLTX flag is now forced regardless of user settings in absence of
layer2 hw acceleration (a6cc0cfa72e0b6d9f2c8fd858aa,
net: Add layer 2 hardware acceleration operations for macvlan devices).

Use netdev_increment_features to rebuild the feature set on capability
changes on either the lower device or on the macvlan interface.

As pointed out by Ben Hutchings, use netdev_update_features on
NETDEV_FEAT_CHANGE event (it calls macvlan_fix_features/netdev_features_change
if needed).

Signed-off-by: Florian Westphal 
Signed-off-by: David S. Miller

Merge branch 'core-locking-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2013-11-14T07:30:30+00:00

Pull core locking changes from Ingo Molnar:
 "The biggest changes:

   - add lockdep support for seqcount/seqlocks structures, this
     unearthed both bugs and required extra annotation.

   - move the various kernel locking primitives to the new
     kernel/locking/ directory"

* 'core-locking-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (21 commits)
  block: Use u64_stats_init() to initialize seqcounts
  locking/lockdep: Mark __lockdep_count_forward_deps() as static
  lockdep/proc: Fix lock-time avg computation
  locking/doc: Update references to kernel/mutex.c
  ipv6: Fix possible ipv6 seqlock deadlock
  cpuset: Fix potential deadlock w/ set_mems_allowed
  seqcount: Add lockdep functionality to seqcount/seqlock structures
  net: Explicitly initialize u64_stats_sync structures for lockdep
  locking: Move the percpu-rwsem code to kernel/locking/
  locking: Move the lglocks code to kernel/locking/
  locking: Move the rwsem code to kernel/locking/
  locking: Move the rtmutex code to kernel/locking/
  locking: Move the semaphore core to kernel/locking/
  locking: Move the spinlock code to kernel/locking/
  locking: Move the lockdep code to kernel/locking/
  locking: Move the mutex code to kernel/locking/
  hung_task debugging: Add tracepoint to report the hang
  x86/locking/kconfig: Update paravirt spinlock Kconfig description
  lockstat: Report avg wait and hold times
  lockdep, x86/alternatives: Drop ancient lockdep fixup message
  ...

net: Add layer 2 hardware acceleration operations for macvlan devices

2013-11-08T00:11:41+00:00

Add a operations structure that allows a network interface to export
the fact that it supports package forwarding in hardware between
physical interfaces and other mac layer devices assigned to it (such
as macvlans). This operaions structure can be used by virtual mac
devices to bypass software switching so that forwarding can be done
in hardware more efficiently.

Signed-off-by: John Fastabend 
Signed-off-by: Neil Horman 
CC: Andy Gospodarek 
CC: "David S. Miller" 
Signed-off-by: David S. Miller

net: Explicitly initialize u64_stats_sync structures for lockdep

2013-11-06T11:40:25+00:00

In order to enable lockdep on seqcount/seqlock structures, we
must explicitly initialize any locks.

The u64_stats_sync structure, uses a seqcount, and thus we need
to introduce a u64_stats_init() function and use it to initialize
the structure.

This unfortunately adds a lot of fairly trivial initialization code
to a number of drivers. But the benefit of ensuring correctness makes
this worth while.

Because these changes are required for lockdep to be enabled, and the
changes are quite trivial, I've not yet split this patch out into 30-some
separate patches, as I figured it would be better to get the various
maintainers thoughts on how to best merge this change along with
the seqcount lockdep enablement.

Feedback would be appreciated!

Signed-off-by: John Stultz 
Acked-by: Julian Anastasov 
Signed-off-by: Peter Zijlstra 
Cc: Alexey Kuznetsov 
Cc: "David S. Miller" 
Cc: Eric Dumazet 
Cc: Hideaki YOSHIFUJI 
Cc: James Morris 
Cc: Jesse Gross 
Cc: Mathieu Desnoyers 
Cc: "Michael S. Tsirkin" 
Cc: Mirko Lindner 
Cc: Patrick McHardy 
Cc: Roger Luethi 
Cc: Rusty Russell 
Cc: Simon Horman 
Cc: Stephen Hemminger 
Cc: Steven Rostedt 
Cc: Thomas Petazzoni 
Cc: Wensong Zhang 
Cc: netdev@vger.kernel.org
Link: http://lkml.kernel.org/r/1381186321-4906-2-git-send-email-john.stultz@linaro.org
Signed-off-by: Ingo Molnar

macvlan: resolve ENOENT errors on creation

2013-10-22T23:22:09+00:00

After the commit below attempting to create macvlan devices was
resulting in ENOENT errors,

# ip link add link p3p2 type macvlan
RTNETLINK answers: Invalid argument

This happens because netdev_upper_dev_link() is called before
register_netdevice() in the macvlan code. Through a call chain
this results in a call to __netdev_adjacent_dev_insert() and
finally a sysfs_create_link(). This requires the kobject of
the macvlan to be registered which is done in register_netdevice().
If there is no kobject which is the case here the ENOENT error
is seen on the command line.

To resolve this move the netdev_upper_dev_link() call below
the register_netdevice() call. This aligns with vlan driver
flow.

Regression introduced here,

commit 5831d66e8097aedfa3bc35941cf265ada2352317
Author: Veaceslav Falico 
Date:   Wed Sep 25 09:20:32 2013 +0200

    net: create sysfs symlinks for neighbour devices

CC: Veaceslav Falico 
CC: Neil Horman 
Signed-off-by: John Fastabend 
Acked-by: Veaceslav Falico 
Acked-by: Neil Horman 
Signed-off-by: David S. Miller

macvlan: Move skb_clone check closer to call

2013-09-11T20:03:12+00:00

Currently macvlan calls skb_clone in macvlan_broadcast but checks
for a NULL return in macvlan_broadcast_one instead.  This is
needlessly confusing and may lead to bugs introduced later.

This patch moves the error check to where the skb_clone call is.

The only other caller of macvlan_broadcast_one never passes in a
NULL value so it doesn't need the check either.

Signed-off-by: Herbert Xu 

Thanks,
Reviewed-by: Simon Horman 
Signed-off-by: David S. Miller

net: macvlan: inherit addr_assign_type along with dev_addr

2013-09-04T00:57:49+00:00

A device inheriting a random or set address should reflect this in
its addr_assign_type.

Cc: Patrick McHardy 
Signed-off-by: Bjørn Mork 
Signed-off-by: David S. Miller