John Hurley says:

====================
Add MPLS actions to TC

This patchset introduces a new TC action module that allows the
manipulation of the MPLS headers of packets. The code impliments
functionality including push, pop, and modify.

Also included are tests for the new funtionality. Note that these will
require iproute2 changes to be submitted soon.

NOTE: these patches are applied to net-next along with the patch:
[PATCH net 1/1] net: openvswitch: fix csum updates for MPLS actions
This patch has been accepted into net but, at time of posting, is not yet
in net-next.

v6-v7:
- add extra tests for setting max/min and exceeding range of fields -
  patch 5 (Roman Mashak)
v5-v6:
- add CONFIG_NET_ACT_MPLS to tc-testing config file - patch 5
  (Davide Caratti)
v4-v5:
- move mpls_hdr() call to after skb_ensure_writable - patch 3
  (Willem de Bruijn)
- move mpls_dec_ttl to helper - patch 4 (Willem de Bruijn)
- add iproute2 usage example to commit msg - patch 4 (David Ahern)
- align label validation with mpls core code - patch 4 (David Ahern)
- improve extack message for no proto in mpls pop - patch 4 (David Ahern)
v3-v4:
- refactor and reuse OvS code (Cong Wang)
- use csum API rather than skb_post*rscum to update skb->csum (Cong Wang)
- remove unnecessary warning (Cong Wang)
- add comments to uapi attributes (David Ahern)
- set strict type policy check for TCA_MPLS_UNSPEC (David Ahern)
- expand/improve extack messages (David Ahern)
- add option to manually set BOS
v2-v3:
- remove a few unnecessary line breaks (Jiri Pirko)
- retract hw offload patch from set (resubmit with driver changes) (Jiri)
v1->v2:
- ensure TCA_ID_MPLS does not conflict with TCA_ID_CTINFO (Davide Caratti)
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Add a new series of selftests to verify the functionality of act_mpls in
TC.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, TC offers the ability to match on the MPLS fields of a packet
through the use of the flow_dissector_key_mpls struct. However, as yet, TC
actions do not allow the modification or manipulation of such fields.

Add a new module that registers TC action ops to allow manipulation of
MPLS. This includes the ability to push and pop headers as well as modify
the contents of new or existing headers. A further action to decrement the
TTL field of an MPLS header is also provided with a new helper added to
support this.

Examples of the usage of the new action with flower rules to push and pop
MPLS labels are:

tc filter add dev eth0 protocol ip parent ffff: flower \
    action mpls push protocol mpls_uc label 123  \
    action mirred egress redirect dev eth1

tc filter add dev eth0 protocol mpls_uc parent ffff: flower \
    action mpls pop protocol ipv4  \
    action mirred egress redirect dev eth1
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Open vSwitch allows the updating of an existing MPLS header on a packet.
In preparation for supporting similar functionality in TC, move this to a
common skb helper function.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Open vSwitch provides code to pop an MPLS header to a packet. In
preparation for supporting this in TC, move the pop code to an skb helper
that can be reused.

Remove the, now unused, update_ethertype static function from OvS.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Open vSwitch provides code to push an MPLS header to a packet. In
preparation for supporting this in TC, move the push code to an skb helper
that can be reused.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Two cases of overlapping changes, nothing fancy.
Signed-off-by: David S. Miller <davem@davemloft.net>

skb_warn_bad_offload and netdev_rx_csum_fault trigger on hard to debug
issues. Dump more state and the header.

Optionally dump the entire packet and linear segment. This is required
to debug checksum bugs that may include bytes past skb_tail_pointer().

Both call sites call this function inside a net_ratelimit() block.
Limit full packet log further to a hard limit of can_dump_full (5).

Based on an earlier patch by Cong Wang, see link below.

Changes v1 -> v2
  - dump frag_list only on full_pkt

Link: https://patchwork.ozlabs.org/patch/1000841/Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Processes can request ipv6 flowlabels with cmsg IPV6_FLOWINFO.
If not set, by default an autogenerated flowlabel is selected.

Explicit flowlabels require a control operation per label plus a
datapath check on every connection (every datagram if unconnected).
This is particularly expensive on unconnected sockets multiplexing
many flows, such as QUIC.

In the common case, where no lease is exclusive, the check can be
safely elided, as both lease request and check trivially succeed.
Indeed, autoflowlabel does the same even with exclusive leases.

Elide the check if no process has requested an exclusive lease.

fl6_sock_lookup previously returns either a reference to a lease or
NULL to denote failure. Modify to return a real error and update
all callers. On return NULL, they can use the label and will elide
the atomic_dec in fl6_sock_release.

This is an optimization. Robust applications still have to revert to
requesting leases if the fast path fails due to an exclusive lease.

Changes RFC->v1:
  - use static_key_false_deferred to rate limit jump label operations
    - call static_key_deferred_flush to stop timers on exit
  - move decrement out of RCU context
  - defer optimization also if opt data is associated with a lease
  - updated all fp6_sock_lookup callers, not just udp
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If an app is playing tricks to reuse a socket via tcp_disconnect(),
bytes_acked/received needs to be reset to 0. Otherwise tcp_info will
report the sum of the current and the old connection..

Cc: Eric Dumazet <edumazet@google.com>
Fixes: 0df48c26 ("tcp: add tcpi_bytes_acked to tcp_info")
Fixes: bdd1f9ed ("tcp: add tcpi_bytes_received to tcp_info")
Signed-off-by: Christoph Paasch <cpaasch@apple.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

IFLA_BOND_PEER_NOTIF_DELAY was set to the value of downdelay instead
of peer_notif_delay. After this change, the correct value is exported.

Fixes: 07a4ddec ("bonding: add an option to specify a delay between peer notifications")
Signed-off-by: Vincent Bernat <vincent@bernat.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

socket->wq is assign-once, set when we are initializing both
struct socket it's in and struct socket_wq it points to.  As the
matter of fact, the only reason for separate allocation was the
ability to RCU-delay freeing of socket_wq.  RCU-delaying the
freeing of socket itself gets rid of that need, so we can just
fold struct socket_wq into the end of struct socket and simplify
the life both for sock_alloc_inode() (one allocation instead of
two) and for tun/tap oddballs, where we used to embed struct socket
and struct socket_wq into the same structure (now - embedding just
the struct socket).

Note that reference to struct socket_wq in struct sock does remain
a reference - that's unchanged.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

we do have an RCU-delayed part there already (freeing the wq),
so it's not like the pipe situation; moreover, it might be
worth considering coallocating wq with the rest of struct sock_alloc.
->sk_wq in struct sock would remain a pointer as it is, but
the object it normally points to would be coallocated with
struct socket...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

Daniel Borkmann says:

====================
pull-request: bpf-next 2019-07-09

The following pull-request contains BPF updates for your *net-next* tree.

The main changes are:

1) Lots of libbpf improvements: i) addition of new APIs to attach BPF
   programs to tracing entities such as {k,u}probes or tracepoints,
   ii) improve specification of BTF-defined maps by eliminating the
   need for data initialization for some of the members, iii) addition
   of a high-level API for setting up and polling perf buffers for
   BPF event output helpers, all from Andrii.

2) Add "prog run" subcommand to bpftool in order to test-run programs
   through the kernel testing infrastructure of BPF, from Quentin.

3) Improve verifier for BPF sockaddr programs to support 8-byte stores
   for user_ip6 and msg_src_ip6 members given clang tends to generate
   such stores, from Stanislav.

4) Enable the new BPF JIT zero-extension optimization for further
   riscv64 ALU ops, from Luke.

5) Fix a bpftool json JIT dump crash on powerpc, from Jiri.

6) Fix an AF_XDP race in generic XDP's receive path, from Ilya.

7) Various smaller fixes from Ilya, Yue and Arnd.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Unlike driver mode, generic xdp receive could be triggered
by different threads on different CPU cores at the same time
leading to the fill and rx queue breakage. For example, this
could happen while sending packets from two processes to the
first interface of veth pair while the second part of it is
open with AF_XDP socket.

Need to take a lock for each generic receive to avoid race.

Fixes: c497176c ("xsk: add Rx receive functions and poll support")
Signed-off-by: Ilya Maximets <i.maximets@samsung.com>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Tested-by: William Tu <u9012063@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

Stephen Suryaputra says:

====================
net: Multipath hashing on inner L3

This series extends commit 363887a2 ("ipv4: Support multipath
hashing on inner IP pkts for GRE tunnel") to include support when the
outer L3 is IPv6 and to consider the case where the inner L3 is
different version from the outer L3, such as IPv6 tunneled by IPv4 GRE
or vice versa. It also includes kselftest scripts to test the use cases.

v2: Clarify the commit messages in the commits in this series to use the
    term tunneled by IPv4 GRE or by IPv6 GRE so that it's clear which
    one is the inner and which one is the outer (per David Miller).
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Add selftest scripts for multipath hashing on inner IP pkts when there
is a single GRE tunnel but there are multiple underlay routes to reach
the other end of the tunnel.

Four cases are covered in these scripts:
    - IPv4 inner, IPv4 outer
    - IPv6 inner, IPv4 outer
    - IPv4 inner, IPv6 outer
    - IPv6 inner, IPv6 outer
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Stephen Suryaputra <ssuryaextr@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make the same support as commit 363887a2 ("ipv4: Support multipath
hashing on inner IP pkts for GRE tunnel") for outer IPv6. The hashing
considers both IPv4 and IPv6 pkts when they are tunneled by IPv6 GRE.
Signed-off-by: Stephen Suryaputra <ssuryaextr@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit 363887a2 ("ipv4: Support multipath hashing on inner IP pkts
for GRE tunnel") supports multipath policy value of 2, Layer 3 or inner
Layer 3 if present, but it only considers inner IPv4. There is a use
case of IPv6 is tunneled by IPv4 GRE, thus add the ability to hash on
inner IPv6 addresses.

Fixes: 363887a2 ("ipv4: Support multipath hashing on inner IP pkts for GRE tunnel")
Signed-off-by: Stephen Suryaputra <ssuryaextr@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The phy_dn variable is still being used in of_phy_connect() after the
of_node_put() call, which may result in use-after-free.

Fixes: 1dd2d06c ("net: Rework pasemi_mac driver to use of_mdio infrastructure")
Signed-off-by: Wen Yang <wen.yang99@zte.com.cn>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: netdev@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>

There is a possible use-after-free issue in the axienet_probe():

1701:	np = of_parse_phandle(pdev->dev.of_node, "axistream-connected", 0);
1702:   if (np) {
...
1787:		of_node_put(np); ---> released here
1788:		lp->eth_irq = platform_get_irq(pdev, 0);
1789:	} else {
...
1801:	}
1802:	if (IS_ERR(lp->dma_regs)) {
...
1805:		of_node_put(np); ---> double released here
1806:		goto free_netdev;
1807:	}

We solve this problem by removing the unnecessary of_node_put().

Fixes: 28ef9ebd ("net: axienet: make use of axistream-connected attribute optional")
Signed-off-by: Wen Yang <wen.yang99@zte.com.cn>
Cc: Anirudha Sarangi <anirudh@xilinx.com>
Cc: John Linn <John.Linn@xilinx.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Michal Simek <michal.simek@xilinx.com>
Cc: Robert Hancock <hancock@sedsystems.ca>
Cc: netdev@vger.kernel.org
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-kernel@vger.kernel.org
Reviewed-by: Robert Hancock <hancock@sedsystems.ca>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fix endianness issue: passing a pointer to 64-bit fd as a 32-bit key
does not work on big-endian architectures. So cast fd to 32-bits when
necessary.
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>

DWMAC4 is capable to support clause 45 mdio communication.
This patch enable the feature on stmmac_mdio_write() and
stmmac_mdio_read() by following phy_write_mmd() and
phy_read_mmd() mdiobus read write implementation format.
Reviewed-by: Li, Yifan <yifan2.li@intel.com>
Signed-off-by: Kweh Hock Leong <hock.leong.kweh@intel.com>
Signed-off-by: Ong Boon Leong <boon.leong.ong@intel.com>
Signed-off-by: Voon Weifeng <weifeng.voon@intel.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Use netif_ovs_is_port() function instead of open code.
This patch doesn't change logic.
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In this release cycle the number of NIC drivers using page_pool
will likely reach 4 drivers.  It is about time to add a maintainer
entry.  Add myself and Ilias.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Acked-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Maxime Chevallier says:

====================
net: mvpp2: Add classification based on the ETHER flow

This series adds support for classification of the ETHER flow in the
mvpp2 driver.

The first patch allows detecting when a user specifies a flow_type that
isn't supported by the driver, while the second adds support for this
flow_type by adding the mapping between the ETHER_FLOW enum value and
the relevant classifier flow entries.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Users can specify classification actions based on the 'ether' flow type.
In that case, this will apply to all ethernet traffic, superseeding
flows such as 'udp4' or 'tcp6'.

Add support for this flow type in the PPv2 classifier, by mapping the
ETHER_FLOW value to the corresponding entries in the classifier.
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add a missing check to detect flow types that we don't support, so that
user can be informed of this.
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If mmap() fails it returns MAP_FAILED, which is defined as ((void *) -1).
The current if-statement incorrectly tests if *ring is NULL.

Fixes: 358be656 ("selftests/net: add txring_overwrite")
Signed-off-by: Frank de Brabander <debrabander@gmail.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Stefano Garzarella says:

====================
vsock/virtio: several fixes in the .probe() and .remove()

During the review of "[PATCH] vsock/virtio: Initialize core virtio vsock
before registering the driver", Stefan pointed out some possible issues
in the .probe() and .remove() callbacks of the virtio-vsock driver.

This series tries to solve these issues:
- Patch 1 adds RCU critical sections to avoid use-after-free of
  'the_virtio_vsock' pointer.
- Patch 2 stops workers before to call vdev->config->reset(vdev) to
  be sure that no one is accessing the device.
- Patch 3 moves the works flush at the end of the .remove() to avoid
  use-after-free of 'vsock' object.

v3:
- Patch 1: use rcu_dereference_protected() to get the_virtio_vosck value in
           the virtio_vsock_probe() [Jason]

v2: https://patchwork.kernel.org/cover/11022343/

v1: https://patchwork.kernel.org/cover/10964733/

Before this series the guest crashes in a few second. After this series the
test runs (~12h) without issues.
Tested on an SMP guest (-smp 4 -monitor tcp:127.0.0.1:1234,server,nowait)
with these scripts to stress the .probe()/.remove() path:

- guest
  while true; do
      cat /dev/urandom | nc-vsock -l 4321 > /dev/null &
      cat /dev/urandom | nc-vsock -l 5321 > /dev/null &
      cat /dev/urandom | nc-vsock -l 6321 > /dev/null &
      cat /dev/urandom | nc-vsock -l 7321 > /dev/null &
      wait
  done

- host
  while true; do
      cat /dev/urandom | nc-vsock 3 4321 > /dev/null &
      cat /dev/urandom | nc-vsock 3 5321 > /dev/null &
      cat /dev/urandom | nc-vsock 3 6321 > /dev/null &
      cat /dev/urandom | nc-vsock 3 7321 > /dev/null &
      sleep 2
      echo "device_del v1" | nc 127.0.0.1 1234
      sleep 1
      echo "device_add vhost-vsock-pci,id=v1,guest-cid=3" | nc 127.0.0.1 1234
      sleep 1
  done
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch moves the flush of works after vdev->config->del_vqs(vdev),
because we need to be sure that no workers run before to free the
'vsock' object.

Since we stopped the workers using the [tx|rx|event]_run flags,
we are sure no one is accessing the device while we are calling
vdev->config->reset(vdev), so we can safely move the workers' flush.

Before the vdev->config->del_vqs(vdev), workers can be scheduled
by VQ callbacks, so we must flush them after del_vqs(), to avoid
use-after-free of 'vsock' object.
Suggested-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Before to call vdev->config->reset(vdev) we need to be sure that
no one is accessing the device, for this reason, we add new variables
in the struct virtio_vsock to stop the workers during the .remove().

This patch also add few comments before vdev->config->reset(vdev)
and vdev->config->del_vqs(vdev).
Suggested-by: Stefan Hajnoczi <stefanha@redhat.com>
Suggested-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Some callbacks used by the upper layers can run while we are in the
.remove(). A potential use-after-free can happen, because we free
the_virtio_vsock without knowing if the callbacks are over or not.

To solve this issue we move the assignment of the_virtio_vsock at the
end of .probe(), when we finished all the initialization, and at the
beginning of .remove(), before to release resources.
For the same reason, we do the same also for the vdev->priv.

We use RCU to be sure that all callbacks that use the_virtio_vsock
ended before freeing it. This is not required for callbacks that
use vdev->priv, because after the vdev->config->del_vqs() we are sure
that they are ended and will no longer be invoked.

We also take the mutex during the .remove() to avoid that .probe() can
run while we are resetting the device.
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Benedikt Spranger says:

====================
Document the configuration of b53

this is the third round to document the configuration of a b53 supported
switch.

v3..v2:
- fix a typo
- improve b53 configuration in DSA_TAG_PROTO_NONE showcase.
- grade up from RFC to patch for mainline inclusion.

v1..v2:
- split out generic parts of the configuration.
- target comments by Andrew Lunn and Florian Fainelli.
- make changes visible to build system
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Document the different needs of documentation for the b53 driver.
Signed-off-by: Benedikt Spranger <b.spranger@linutronix.de>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Document DSA tagged and VLAN based switch configuration by showcases.
Signed-off-by: Benedikt Spranger <b.spranger@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fix to return negative error code -EINVAL from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 1f35a56c ("nfp: tls: add/delete TLS TX connections")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Michael Chan says:

====================
bnxt_en: Add XDP_REDIRECT support.

This patch series adds XDP_REDIRECT support by Andy Gospodarek.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This removes contention over page allocation for XDP_REDIRECT actions by
adding page_pool support per queue for the driver.  The performance for
XDP_REDIRECT actions scales linearly with the number of cores performing
redirect actions when using the page pools instead of the standard page
allocator.

v2: Fix up the error path from XDP registration, noted by Ilias Apalodimas.
Signed-off-by: Andy Gospodarek <gospo@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This adds basic support for XDP_REDIRECT in the bnxt_en driver.  Next
patch adds the more optimized page pool support.
Signed-off-by: Andy Gospodarek <gospo@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>