[PATCH net-next] veth: set iflink to the peer veth

[PATCH net-next] veth: set iflink to the peer veth

[Nicolas Dichtel]

Now that the peer netns is advertised in rtnl messages, we can set this property
so that IFLA_LINK will advertise the peer ifindex. It allows the userland to get
the full veth configuration.

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 drivers/net/veth.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/veth.c b/drivers/net/veth.c
index 4cca36ebc4fb..02d1f798a371 100644
--- a/drivers/net/veth.c
+++ b/drivers/net/veth.c
@@ -427,9 +427,11 @@ static int veth_newlink(struct net *src_net, struct net_device *dev,
 
 	priv = netdev_priv(dev);
 	rcu_assign_pointer(priv->peer, peer);
+	dev->iflink = peer->ifindex;
 
 	priv = netdev_priv(peer);
 	rcu_assign_pointer(priv->peer, dev);
+	peer->iflink = dev->ifindex;
 	return 0;
 
 err_register_dev:
-- 
2.2.2

Re: [PATCH net-next] veth: set iflink to the peer veth

[Eric Dumazet]

On Tue, 2015-02-03 at 16:41 +0100, Nicolas Dichtel wrote:
> Now that the peer netns is advertised in rtnl messages, we can set this property
> so that IFLA_LINK will advertise the peer ifindex. It allows the userland to get
> the full veth configuration.
> 
> Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
> ---
>  drivers/net/veth.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/net/veth.c b/drivers/net/veth.c
> index 4cca36ebc4fb..02d1f798a371 100644
> --- a/drivers/net/veth.c
> +++ b/drivers/net/veth.c
> @@ -427,9 +427,11 @@ static int veth_newlink(struct net *src_net, struct net_device *dev,
>  
>  	priv = netdev_priv(dev);
>  	rcu_assign_pointer(priv->peer, peer);
> +	dev->iflink = peer->ifindex;
>  
>  	priv = netdev_priv(peer);
>  	rcu_assign_pointer(priv->peer, dev);
> +	peer->iflink = dev->ifindex;
>  	return 0;
>  
>  err_register_dev:

Is is network namespace ready ?

If I move one veth to another namespace, is iflink still relevant ?

Re: [PATCH net-next] veth: set iflink to the peer veth

[Nicolas Dichtel]

Le 03/02/2015 17:12, Eric Dumazet a écrit :
> If I move one veth to another namespace, is iflink still relevant ?
>
>
You're right. ifindex may change.

[PATCH net-next v2 0/12] Remove iflink field from the net_device structure

[Nicolas Dichtel]

The first goal of this series was to advertise the veth peer via the IFLA_LINK
attribute, but iflink was not ready for network namespaces.

The iflink of an interface should be set to its ifindex for a physical interface
and to another value (0 if not relevant) for a virtual interface.
This was not the case for some interfaces, like vxlan, bond, or bridge for
example.
There is also a risk, if the targeted interface moves to another netns, that the
ifindex changes without updating corresponding iflink fields (eg. vlan).

Moving the management of this property into virtual interface drivers allows to
better handle this last case because most of virtual interface drivers have a
pointer to the link netdevice.
Anyway, dev->iflink value was always a copy of some internal data of the virtual
interface driver, thus let's use these internal data directly.

So, this series removes the iflink field and let the drivers manage it.
Only the last patch was present in the v1, but I fully rework it.

 drivers/infiniband/ulp/ipoib/ipoib_main.c |  8 ++++++++
 drivers/infiniband/ulp/ipoib/ipoib_vlan.c |  1 -
 drivers/net/ipvlan/ipvlan_core.c          |  2 +-
 drivers/net/ipvlan/ipvlan_main.c          |  9 ++++++++-
 drivers/net/macvlan.c                     |  9 ++++++++-
 drivers/net/veth.c                        | 15 +++++++++++++++
 include/linux/netdevice.h                 |  7 +++++--
 include/net/ip6_tunnel.h                  |  1 +
 include/net/ip_tunnels.h                  |  1 +
 net/8021q/vlan_dev.c                      |  9 ++++++++-
 net/batman-adv/hard-interface.c           |  5 +++--
 net/bridge/br_netlink.c                   |  4 ++--
 net/core/dev.c                            | 32 +++++++++++++++++++++----------
 net/core/link_watch.c                     |  4 ++--
 net/core/net-sysfs.c                      | 10 +++++++++-
 net/core/rtnetlink.c                      |  8 ++++----
 net/dsa/slave.c                           |  8 +++-----
 net/ipv4/ip_gre.c                         |  2 ++
 net/ipv4/ip_tunnel.c                      |  9 ++++++++-
 net/ipv4/ip_vti.c                         |  2 +-
 net/ipv4/ipip.c                           |  2 +-
 net/ipv4/ipmr.c                           |  9 +++++++--
 net/ipv6/addrconf.c                       |  4 ++--
 net/ipv6/ip6_gre.c                        |  8 ++------
 net/ipv6/ip6_tunnel.c                     | 10 ++++++++--
 net/ipv6/ip6_vti.c                        |  3 +--
 net/ipv6/ip6mr.c                          |  9 +++++++--
 net/ipv6/sit.c                            |  3 +--
 28 files changed, 140 insertions(+), 54 deletions(-)

Comments are welcome.

Regards,
Nicolas

[PATCH net-next v2 01/12] dev: introduce dev_get_iflink()

[Nicolas Dichtel]

The goal of this patch is to prepare the removal of the iflink field. It
introduces a new ndo function, which will be implemented by virtual interfaces.

There is no functional change into this patch. All readers of iflink field
now call dev_get_iflink().

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 drivers/net/ipvlan/ipvlan_core.c |  2 +-
 include/linux/netdevice.h        |  4 ++++
 net/batman-adv/hard-interface.c  |  5 +++--
 net/bridge/br_netlink.c          |  4 ++--
 net/core/dev.c                   | 21 +++++++++++++++++++--
 net/core/link_watch.c            |  4 ++--
 net/core/net-sysfs.c             | 10 +++++++++-
 net/core/rtnetlink.c             |  8 ++++----
 net/ipv4/ipmr.c                  |  2 +-
 net/ipv6/addrconf.c              |  4 ++--
 net/ipv6/ip6mr.c                 |  2 +-
 11 files changed, 48 insertions(+), 18 deletions(-)

diff --git a/drivers/net/ipvlan/ipvlan_core.c b/drivers/net/ipvlan/ipvlan_core.c
index 2a175006028b..131bde98188d 100644
--- a/drivers/net/ipvlan/ipvlan_core.c
+++ b/drivers/net/ipvlan/ipvlan_core.c
@@ -330,7 +330,7 @@ static int ipvlan_process_v4_outbound(struct sk_buff *skb)
 	struct rtable *rt;
 	int err, ret = NET_XMIT_DROP;
 	struct flowi4 fl4 = {
-		.flowi4_oif = dev->iflink,
+		.flowi4_oif = dev_get_iflink(dev),
 		.flowi4_tos = RT_TOS(ip4h->tos),
 		.flowi4_flags = FLOWI_FLAG_ANYSRC,
 		.daddr = ip4h->daddr,
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 967bb4c8caf1..788eb7a622ad 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -1030,6 +1030,8 @@ typedef u16 (*select_queue_fallback_t)(struct net_device *dev,
  *			     int queue_index, u32 maxrate);
  *	Called when a user wants to set a max-rate limitation of specific
  *	TX queue.
+ * int (*ndo_get_iflink)(const struct net_device *dev);
+ *	Called to get the iflink value of this device.
  */
 struct net_device_ops {
 	int			(*ndo_init)(struct net_device *dev);
@@ -1191,6 +1193,7 @@ struct net_device_ops {
 	int			(*ndo_set_tx_maxrate)(struct net_device *dev,
 						      int queue_index,
 						      u32 maxrate);
+	int			(*ndo_get_iflink)(const struct net_device *dev);
 };
 
 /**
@@ -2149,6 +2152,7 @@ void __dev_remove_pack(struct packet_type *pt);
 void dev_add_offload(struct packet_offload *po);
 void dev_remove_offload(struct packet_offload *po);
 
+int dev_get_iflink(const struct net_device *dev);
 struct net_device *__dev_get_by_flags(struct net *net, unsigned short flags,
 				      unsigned short mask);
 struct net_device *dev_get_by_name(struct net *net, const char *name);
diff --git a/net/batman-adv/hard-interface.c b/net/batman-adv/hard-interface.c
index fbda6b54baff..baf1f9843f2c 100644
--- a/net/batman-adv/hard-interface.c
+++ b/net/batman-adv/hard-interface.c
@@ -83,11 +83,12 @@ static bool batadv_is_on_batman_iface(const struct net_device *net_dev)
 		return true;
 
 	/* no more parents..stop recursion */
-	if (net_dev->iflink == 0 || net_dev->iflink == net_dev->ifindex)
+	if (dev_get_iflink(net_dev) == 0 ||
+	    dev_get_iflink(net_dev) == net_dev->ifindex)
 		return false;
 
 	/* recurse over the parent device */
-	parent_dev = __dev_get_by_index(&init_net, net_dev->iflink);
+	parent_dev = __dev_get_by_index(&init_net, dev_get_iflink(net_dev));
 	/* if we got a NULL parent_dev there is something broken.. */
 	if (WARN(!parent_dev, "Cannot find parent device"))
 		return false;
diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c
index e1115a224a95..0e4ddb81610d 100644
--- a/net/bridge/br_netlink.c
+++ b/net/bridge/br_netlink.c
@@ -305,8 +305,8 @@ static int br_fill_ifinfo(struct sk_buff *skb,
 	    nla_put_u8(skb, IFLA_OPERSTATE, operstate) ||
 	    (dev->addr_len &&
 	     nla_put(skb, IFLA_ADDRESS, dev->addr_len, dev->dev_addr)) ||
-	    (dev->ifindex != dev->iflink &&
-	     nla_put_u32(skb, IFLA_LINK, dev->iflink)))
+	    (dev->ifindex != dev_get_iflink(dev) &&
+	     nla_put_u32(skb, IFLA_LINK, dev_get_iflink(dev))))
 		goto nla_put_failure;
 
 	if (event == RTM_NEWLINK && port) {
diff --git a/net/core/dev.c b/net/core/dev.c
index 65492b0354c0..77172d085760 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -660,6 +660,23 @@ __setup("netdev=", netdev_boot_setup);
 *******************************************************************************/
 
 /**
+ *	dev_get_iflink	- get 'iflink' value of a interface
+ *	@dev: targeted interface
+ *
+ *	Indicates the ifindex the interface is linked to.
+ *	Physical interfaces have the same 'ifindex' and 'iflink' values.
+ */
+
+int dev_get_iflink(const struct net_device *dev)
+{
+	if (dev->netdev_ops && dev->netdev_ops->ndo_get_iflink)
+		return dev->netdev_ops->ndo_get_iflink(dev);
+
+	return dev->iflink;
+}
+EXPORT_SYMBOL(dev_get_iflink);
+
+/**
  *	__dev_get_by_name	- find a device by its name
  *	@net: the applicable net namespace
  *	@name: name to find
@@ -6345,7 +6362,7 @@ int register_netdevice(struct net_device *dev)
 	else if (__dev_get_by_index(net, dev->ifindex))
 		goto err_uninit;
 
-	if (dev->iflink == -1)
+	if (dev_get_iflink(dev) == -1)
 		dev->iflink = dev->ifindex;
 
 	/* Transfer changeable features to wanted_features and enable
@@ -7061,7 +7078,7 @@ int dev_change_net_namespace(struct net_device *dev, struct net *net, const char
 
 	/* If there is an ifindex conflict assign a new one */
 	if (__dev_get_by_index(net, dev->ifindex)) {
-		int iflink = (dev->iflink == dev->ifindex);
+		int iflink = (dev_get_iflink(dev) == dev->ifindex);
 		dev->ifindex = dev_new_index(net);
 		if (iflink)
 			dev->iflink = dev->ifindex;
diff --git a/net/core/link_watch.c b/net/core/link_watch.c
index 49a9e3e06c08..982861607f88 100644
--- a/net/core/link_watch.c
+++ b/net/core/link_watch.c
@@ -40,7 +40,7 @@ static DEFINE_SPINLOCK(lweventlist_lock);
 static unsigned char default_operstate(const struct net_device *dev)
 {
 	if (!netif_carrier_ok(dev))
-		return (dev->ifindex != dev->iflink ?
+		return (dev->ifindex != dev_get_iflink(dev) ?
 			IF_OPER_LOWERLAYERDOWN : IF_OPER_DOWN);
 
 	if (netif_dormant(dev))
@@ -89,7 +89,7 @@ static bool linkwatch_urgent_event(struct net_device *dev)
 	if (!netif_running(dev))
 		return false;
 
-	if (dev->ifindex != dev->iflink)
+	if (dev->ifindex != dev_get_iflink(dev))
 		return true;
 
 	if (dev->priv_flags & IFF_TEAM_PORT)
diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c
index cc5cf689809c..4238d6da5c60 100644
--- a/net/core/net-sysfs.c
+++ b/net/core/net-sysfs.c
@@ -109,11 +109,19 @@ NETDEVICE_SHOW_RO(dev_id, fmt_hex);
 NETDEVICE_SHOW_RO(dev_port, fmt_dec);
 NETDEVICE_SHOW_RO(addr_assign_type, fmt_dec);
 NETDEVICE_SHOW_RO(addr_len, fmt_dec);
-NETDEVICE_SHOW_RO(iflink, fmt_dec);
 NETDEVICE_SHOW_RO(ifindex, fmt_dec);
 NETDEVICE_SHOW_RO(type, fmt_dec);
 NETDEVICE_SHOW_RO(link_mode, fmt_dec);
 
+static ssize_t iflink_show(struct device *dev, struct device_attribute *attr,
+			   char *buf)
+{
+	struct net_device *ndev = to_net_dev(dev);
+
+	return sprintf(buf, fmt_dec, dev_get_iflink(ndev));
+}
+static DEVICE_ATTR_RO(iflink);
+
 static ssize_t format_name_assign_type(const struct net_device *dev, char *buf)
 {
 	return sprintf(buf, fmt_dec, dev->name_assign_type);
diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
index b96ac2109c82..ee0186cdd5cf 100644
--- a/net/core/rtnetlink.c
+++ b/net/core/rtnetlink.c
@@ -1055,8 +1055,8 @@ static int rtnl_fill_ifinfo(struct sk_buff *skb, struct net_device *dev,
 #ifdef CONFIG_RPS
 	    nla_put_u32(skb, IFLA_NUM_RX_QUEUES, dev->num_rx_queues) ||
 #endif
-	    (dev->ifindex != dev->iflink &&
-	     nla_put_u32(skb, IFLA_LINK, dev->iflink)) ||
+	    (dev->ifindex != dev_get_iflink(dev) &&
+	     nla_put_u32(skb, IFLA_LINK, dev_get_iflink(dev))) ||
 	    (upper_dev &&
 	     nla_put_u32(skb, IFLA_MASTER, upper_dev->ifindex)) ||
 	    nla_put_u8(skb, IFLA_CARRIER, netif_carrier_ok(dev)) ||
@@ -2863,8 +2863,8 @@ int ndo_dflt_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq,
 	     nla_put_u32(skb, IFLA_MASTER, br_dev->ifindex)) ||
 	    (dev->addr_len &&
 	     nla_put(skb, IFLA_ADDRESS, dev->addr_len, dev->dev_addr)) ||
-	    (dev->ifindex != dev->iflink &&
-	     nla_put_u32(skb, IFLA_LINK, dev->iflink)))
+	    (dev->ifindex != dev_get_iflink(dev) &&
+	     nla_put_u32(skb, IFLA_LINK, dev_get_iflink(dev))))
 		goto nla_put_failure;
 
 	br_afspec = nla_nest_start(skb, IFLA_AF_SPEC);
diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c
index b4a545d24adb..eec68b0c3bc8 100644
--- a/net/ipv4/ipmr.c
+++ b/net/ipv4/ipmr.c
@@ -801,7 +801,7 @@ static int vif_add(struct net *net, struct mr_table *mrt,
 	v->pkt_out = 0;
 	v->link = dev->ifindex;
 	if (v->flags & (VIFF_TUNNEL | VIFF_REGISTER))
-		v->link = dev->iflink;
+		v->link = dev_get_iflink(dev);
 
 	/* And finish update writing critical data */
 	write_lock_bh(&mrt_lock);
diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index 5c9e94cb1b2c..37b70e82bff8 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -4858,8 +4858,8 @@ static int inet6_fill_ifinfo(struct sk_buff *skb, struct inet6_dev *idev,
 	    (dev->addr_len &&
 	     nla_put(skb, IFLA_ADDRESS, dev->addr_len, dev->dev_addr)) ||
 	    nla_put_u32(skb, IFLA_MTU, dev->mtu) ||
-	    (dev->ifindex != dev->iflink &&
-	     nla_put_u32(skb, IFLA_LINK, dev->iflink)))
+	    (dev->ifindex != dev_get_iflink(dev) &&
+	     nla_put_u32(skb, IFLA_LINK, dev_get_iflink(dev))))
 		goto nla_put_failure;
 	protoinfo = nla_nest_start(skb, IFLA_PROTINFO);
 	if (!protoinfo)
diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c
index caf6b99374e6..18a5ab286420 100644
--- a/net/ipv6/ip6mr.c
+++ b/net/ipv6/ip6mr.c
@@ -992,7 +992,7 @@ static int mif6_add(struct net *net, struct mr6_table *mrt,
 	v->pkt_out = 0;
 	v->link = dev->ifindex;
 	if (v->flags & MIFF_REGISTER)
-		v->link = dev->iflink;
+		v->link = dev_get_iflink(dev);
 
 	/* And finish update writing critical data */
 	write_lock_bh(&mrt_lock);
-- 
2.2.2

[PATCH net-next v2 02/12] ip6tnl,gre6,vti6: implement ndo_get_iflink

[Nicolas Dichtel]

Don't use dev->iflink anymore.

CC: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 include/net/ip6_tunnel.h |  1 +
 net/ipv6/ip6_gre.c       |  8 ++------
 net/ipv6/ip6_tunnel.c    | 10 ++++++++--
 net/ipv6/ip6_vti.c       |  3 +--
 4 files changed, 12 insertions(+), 10 deletions(-)

diff --git a/include/net/ip6_tunnel.h b/include/net/ip6_tunnel.h
index 76c091b53dae..1668be5937e6 100644
--- a/include/net/ip6_tunnel.h
+++ b/include/net/ip6_tunnel.h
@@ -71,6 +71,7 @@ __u16 ip6_tnl_parse_tlv_enc_lim(struct sk_buff *skb, __u8 *raw);
 __u32 ip6_tnl_get_cap(struct ip6_tnl *t, const struct in6_addr *laddr,
 			     const struct in6_addr *raddr);
 struct net *ip6_tnl_get_link_net(const struct net_device *dev);
+int ip6_tnl_get_iflink(const struct net_device *dev);
 
 static inline void ip6tunnel_xmit(struct sk_buff *skb, struct net_device *dev)
 {
diff --git a/net/ipv6/ip6_gre.c b/net/ipv6/ip6_gre.c
index 0f4e73da14e4..f724329d7436 100644
--- a/net/ipv6/ip6_gre.c
+++ b/net/ipv6/ip6_gre.c
@@ -1216,6 +1216,7 @@ static const struct net_device_ops ip6gre_netdev_ops = {
 	.ndo_do_ioctl		= ip6gre_tunnel_ioctl,
 	.ndo_change_mtu		= ip6gre_tunnel_change_mtu,
 	.ndo_get_stats64	= ip_tunnel_get_stats64,
+	.ndo_get_iflink		= ip6_tnl_get_iflink,
 };
 
 static void ip6gre_dev_free(struct net_device *dev)
@@ -1238,7 +1239,6 @@ static void ip6gre_tunnel_setup(struct net_device *dev)
 	if (!(t->parms.flags & IP6_TNL_F_IGN_ENCAP_LIMIT))
 		dev->mtu -= 8;
 	dev->flags |= IFF_NOARP;
-	dev->iflink = 0;
 	dev->addr_len = sizeof(struct in6_addr);
 	netif_keep_dst(dev);
 }
@@ -1270,8 +1270,6 @@ static int ip6gre_tunnel_init(struct net_device *dev)
 		u64_stats_init(&ip6gre_tunnel_stats->syncp);
 	}
 
-	dev->iflink = tunnel->parms.link;
-
 	return 0;
 }
 
@@ -1480,8 +1478,6 @@ static int ip6gre_tap_init(struct net_device *dev)
 	if (!dev->tstats)
 		return -ENOMEM;
 
-	dev->iflink = tunnel->parms.link;
-
 	return 0;
 }
 
@@ -1493,6 +1489,7 @@ static const struct net_device_ops ip6gre_tap_netdev_ops = {
 	.ndo_validate_addr = eth_validate_addr,
 	.ndo_change_mtu = ip6gre_tunnel_change_mtu,
 	.ndo_get_stats64 = ip_tunnel_get_stats64,
+	.ndo_get_iflink = ip6_tnl_get_iflink,
 };
 
 static void ip6gre_tap_setup(struct net_device *dev)
@@ -1503,7 +1500,6 @@ static void ip6gre_tap_setup(struct net_device *dev)
 	dev->netdev_ops = &ip6gre_tap_netdev_ops;
 	dev->destructor = ip6gre_dev_free;
 
-	dev->iflink = 0;
 	dev->features |= NETIF_F_NETNS_LOCAL;
 }
 
diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
index 9bd85f0dff69..b6a211a150b2 100644
--- a/net/ipv6/ip6_tunnel.c
+++ b/net/ipv6/ip6_tunnel.c
@@ -1264,8 +1264,6 @@ static void ip6_tnl_link_config(struct ip6_tnl *t)
 	else
 		dev->flags &= ~IFF_POINTOPOINT;
 
-	dev->iflink = p->link;
-
 	if (p->flags & IP6_TNL_F_CAP_XMIT) {
 		int strict = (ipv6_addr_type(&p->raddr) &
 			      (IPV6_ADDR_MULTICAST|IPV6_ADDR_LINKLOCAL));
@@ -1517,6 +1515,13 @@ ip6_tnl_change_mtu(struct net_device *dev, int new_mtu)
 	return 0;
 }
 
+int ip6_tnl_get_iflink(const struct net_device *dev)
+{
+	struct ip6_tnl *t = netdev_priv(dev);
+
+	return t->parms.link;
+}
+EXPORT_SYMBOL(ip6_tnl_get_iflink);
 
 static const struct net_device_ops ip6_tnl_netdev_ops = {
 	.ndo_init	= ip6_tnl_dev_init,
@@ -1525,6 +1530,7 @@ static const struct net_device_ops ip6_tnl_netdev_ops = {
 	.ndo_do_ioctl	= ip6_tnl_ioctl,
 	.ndo_change_mtu = ip6_tnl_change_mtu,
 	.ndo_get_stats	= ip6_get_stats,
+	.ndo_get_iflink = ip6_tnl_get_iflink,
 };
 
 
diff --git a/net/ipv6/ip6_vti.c b/net/ipv6/ip6_vti.c
index 53d90ed68905..b53148444e15 100644
--- a/net/ipv6/ip6_vti.c
+++ b/net/ipv6/ip6_vti.c
@@ -601,8 +601,6 @@ static void vti6_link_config(struct ip6_tnl *t)
 		dev->flags |= IFF_POINTOPOINT;
 	else
 		dev->flags &= ~IFF_POINTOPOINT;
-
-	dev->iflink = p->link;
 }
 
 /**
@@ -808,6 +806,7 @@ static const struct net_device_ops vti6_netdev_ops = {
 	.ndo_do_ioctl	= vti6_ioctl,
 	.ndo_change_mtu = vti6_change_mtu,
 	.ndo_get_stats64 = ip_tunnel_get_stats64,
+	.ndo_get_iflink = ip6_tnl_get_iflink,
 };
 
 /**
-- 
2.2.2

[PATCH net-next v2 03/12] ipip,gre,vti,sit: implement ndo_get_iflink

[Nicolas Dichtel]

Don't use dev->iflink anymore.

CC: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 include/net/ip_tunnels.h | 1 +
 net/ipv4/ip_gre.c        | 2 ++
 net/ipv4/ip_tunnel.c     | 9 ++++++++-
 net/ipv4/ip_vti.c        | 2 +-
 net/ipv4/ipip.c          | 2 +-
 net/ipv6/sit.c           | 3 +--
 6 files changed, 14 insertions(+), 5 deletions(-)

diff --git a/include/net/ip_tunnels.h b/include/net/ip_tunnels.h
index 2c47061a6954..d8214cb88bbc 100644
--- a/include/net/ip_tunnels.h
+++ b/include/net/ip_tunnels.h
@@ -142,6 +142,7 @@ int ip_tunnel_init(struct net_device *dev);
 void ip_tunnel_uninit(struct net_device *dev);
 void  ip_tunnel_dellink(struct net_device *dev, struct list_head *head);
 struct net *ip_tunnel_get_link_net(const struct net_device *dev);
+int ip_tunnel_get_iflink(const struct net_device *dev);
 int ip_tunnel_init_net(struct net *net, int ip_tnl_net_id,
 		       struct rtnl_link_ops *ops, char *devname);
 
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
index 0eb2a040a830..1060ca0bc23a 100644
--- a/net/ipv4/ip_gre.c
+++ b/net/ipv4/ip_gre.c
@@ -456,6 +456,7 @@ static const struct net_device_ops ipgre_netdev_ops = {
 	.ndo_do_ioctl		= ipgre_tunnel_ioctl,
 	.ndo_change_mtu		= ip_tunnel_change_mtu,
 	.ndo_get_stats64	= ip_tunnel_get_stats64,
+	.ndo_get_iflink		= ip_tunnel_get_iflink,
 };
 
 #define GRE_FEATURES (NETIF_F_SG |		\
@@ -686,6 +687,7 @@ static const struct net_device_ops gre_tap_netdev_ops = {
 	.ndo_validate_addr	= eth_validate_addr,
 	.ndo_change_mtu		= ip_tunnel_change_mtu,
 	.ndo_get_stats64	= ip_tunnel_get_stats64,
+	.ndo_get_iflink		= ip_tunnel_get_iflink,
 };
 
 static void ipgre_tap_setup(struct net_device *dev)
diff --git a/net/ipv4/ip_tunnel.c b/net/ipv4/ip_tunnel.c
index 2cd08280c77b..4bb7252110a6 100644
--- a/net/ipv4/ip_tunnel.c
+++ b/net/ipv4/ip_tunnel.c
@@ -389,7 +389,6 @@ static int ip_tunnel_bind_dev(struct net_device *dev)
 		hlen = tdev->hard_header_len + tdev->needed_headroom;
 		mtu = tdev->mtu;
 	}
-	dev->iflink = tunnel->parms.link;
 
 	dev->needed_headroom = t_hlen + hlen;
 	mtu -= (dev->hard_header_len + t_hlen);
@@ -980,6 +979,14 @@ struct net *ip_tunnel_get_link_net(const struct net_device *dev)
 }
 EXPORT_SYMBOL(ip_tunnel_get_link_net);
 
+int ip_tunnel_get_iflink(const struct net_device *dev)
+{
+	struct ip_tunnel *tunnel = netdev_priv(dev);
+
+	return tunnel->parms.link;
+}
+EXPORT_SYMBOL(ip_tunnel_get_iflink);
+
 int ip_tunnel_init_net(struct net *net, int ip_tnl_net_id,
 				  struct rtnl_link_ops *ops, char *devname)
 {
diff --git a/net/ipv4/ip_vti.c b/net/ipv4/ip_vti.c
index 5a6e27054f0a..c4f93c0d1104 100644
--- a/net/ipv4/ip_vti.c
+++ b/net/ipv4/ip_vti.c
@@ -341,6 +341,7 @@ static const struct net_device_ops vti_netdev_ops = {
 	.ndo_do_ioctl	= vti_tunnel_ioctl,
 	.ndo_change_mtu	= ip_tunnel_change_mtu,
 	.ndo_get_stats64 = ip_tunnel_get_stats64,
+	.ndo_get_iflink = ip_tunnel_get_iflink,
 };
 
 static void vti_tunnel_setup(struct net_device *dev)
@@ -361,7 +362,6 @@ static int vti_tunnel_init(struct net_device *dev)
 	dev->hard_header_len	= LL_MAX_HEADER + sizeof(struct iphdr);
 	dev->mtu		= ETH_DATA_LEN;
 	dev->flags		= IFF_NOARP;
-	dev->iflink		= 0;
 	dev->addr_len		= 4;
 	dev->features		|= NETIF_F_LLTX;
 	netif_keep_dst(dev);
diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c
index bfbcc85c02ee..5c81f6e40842 100644
--- a/net/ipv4/ipip.c
+++ b/net/ipv4/ipip.c
@@ -272,6 +272,7 @@ static const struct net_device_ops ipip_netdev_ops = {
 	.ndo_do_ioctl	= ipip_tunnel_ioctl,
 	.ndo_change_mtu = ip_tunnel_change_mtu,
 	.ndo_get_stats64 = ip_tunnel_get_stats64,
+	.ndo_get_iflink = ip_tunnel_get_iflink,
 };
 
 #define IPIP_FEATURES (NETIF_F_SG |		\
@@ -286,7 +287,6 @@ static void ipip_tunnel_setup(struct net_device *dev)
 
 	dev->type		= ARPHRD_TUNNEL;
 	dev->flags		= IFF_NOARP;
-	dev->iflink		= 0;
 	dev->addr_len		= 4;
 	dev->features		|= NETIF_F_LLTX;
 	netif_keep_dst(dev);
diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c
index e6b9f51b15e8..6cf2026a9cea 100644
--- a/net/ipv6/sit.c
+++ b/net/ipv6/sit.c
@@ -1076,7 +1076,6 @@ static void ipip6_tunnel_bind_dev(struct net_device *dev)
 		if (dev->mtu < IPV6_MIN_MTU)
 			dev->mtu = IPV6_MIN_MTU;
 	}
-	dev->iflink = tunnel->parms.link;
 }
 
 static void ipip6_tunnel_update(struct ip_tunnel *t, struct ip_tunnel_parm *p)
@@ -1336,6 +1335,7 @@ static const struct net_device_ops ipip6_netdev_ops = {
 	.ndo_do_ioctl	= ipip6_tunnel_ioctl,
 	.ndo_change_mtu	= ipip6_tunnel_change_mtu,
 	.ndo_get_stats64 = ip_tunnel_get_stats64,
+	.ndo_get_iflink = ip_tunnel_get_iflink,
 };
 
 static void ipip6_dev_free(struct net_device *dev)
@@ -1366,7 +1366,6 @@ static void ipip6_tunnel_setup(struct net_device *dev)
 	dev->mtu		= ETH_DATA_LEN - t_hlen;
 	dev->flags		= IFF_NOARP;
 	netif_keep_dst(dev);
-	dev->iflink		= 0;
 	dev->addr_len		= 4;
 	dev->features		|= NETIF_F_LLTX;
 	dev->features		|= SIT_FEATURES;
-- 
2.2.2

[PATCH net-next v2 04/12] ipmr,ip6mr: implement ndo_get_iflink

[Nicolas Dichtel]

Don't use dev->iflink anymore.

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 net/ipv4/ipmr.c  | 7 ++++++-
 net/ipv6/ip6mr.c | 7 ++++++-
 2 files changed, 12 insertions(+), 2 deletions(-)

diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c
index eec68b0c3bc8..d2e3b3ef039e 100644
--- a/net/ipv4/ipmr.c
+++ b/net/ipv4/ipmr.c
@@ -473,8 +473,14 @@ static netdev_tx_t reg_vif_xmit(struct sk_buff *skb, struct net_device *dev)
 	return NETDEV_TX_OK;
 }
 
+static int reg_vif_get_iflink(const struct net_device *dev)
+{
+	return 0;
+}
+
 static const struct net_device_ops reg_vif_netdev_ops = {
 	.ndo_start_xmit	= reg_vif_xmit,
+	.ndo_get_iflink = reg_vif_get_iflink,
 };
 
 static void reg_vif_setup(struct net_device *dev)
@@ -509,7 +515,6 @@ static struct net_device *ipmr_reg_vif(struct net *net, struct mr_table *mrt)
 		free_netdev(dev);
 		return NULL;
 	}
-	dev->iflink = 0;
 
 	rcu_read_lock();
 	in_dev = __in_dev_get_rcu(dev);
diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c
index 18a5ab286420..003431f5b4b6 100644
--- a/net/ipv6/ip6mr.c
+++ b/net/ipv6/ip6mr.c
@@ -718,8 +718,14 @@ static netdev_tx_t reg_vif_xmit(struct sk_buff *skb,
 	return NETDEV_TX_OK;
 }
 
+static int reg_vif_get_iflink(const struct net_device *dev)
+{
+	return 0;
+}
+
 static const struct net_device_ops reg_vif_netdev_ops = {
 	.ndo_start_xmit	= reg_vif_xmit,
+	.ndo_get_iflink = reg_vif_get_iflink,
 };
 
 static void reg_vif_setup(struct net_device *dev)
@@ -752,7 +758,6 @@ static struct net_device *ip6mr_reg_vif(struct net *net, struct mr6_table *mrt)
 		free_netdev(dev);
 		return NULL;
 	}
-	dev->iflink = 0;
 
 	if (dev_open(dev))
 		goto failure;
-- 
2.2.2

[PATCH net-next v2 05/12] vlan: implement ndo_get_iflink

[Nicolas Dichtel]

Don't use dev->iflink anymore.

CC: Patrick McHardy <kaber@trash.net>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 net/8021q/vlan_dev.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c
index 8b5ab9033b41..01d7ba840df8 100644
--- a/net/8021q/vlan_dev.c
+++ b/net/8021q/vlan_dev.c
@@ -538,7 +538,6 @@ static int vlan_dev_init(struct net_device *dev)
 	/* IFF_BROADCAST|IFF_MULTICAST; ??? */
 	dev->flags  = real_dev->flags & ~(IFF_UP | IFF_PROMISC | IFF_ALLMULTI |
 					  IFF_MASTER | IFF_SLAVE);
-	dev->iflink = real_dev->ifindex;
 	dev->state  = (real_dev->state & ((1<<__LINK_STATE_NOCARRIER) |
 					  (1<<__LINK_STATE_DORMANT))) |
 		      (1<<__LINK_STATE_PRESENT);
@@ -733,6 +732,13 @@ static void vlan_dev_netpoll_cleanup(struct net_device *dev)
 }
 #endif /* CONFIG_NET_POLL_CONTROLLER */
 
+static int vlan_dev_get_iflink(const struct net_device *dev)
+{
+	struct net_device *real_dev = vlan_dev_priv(dev)->real_dev;
+
+	return real_dev->ifindex;
+}
+
 static const struct ethtool_ops vlan_ethtool_ops = {
 	.get_settings	        = vlan_ethtool_get_settings,
 	.get_drvinfo	        = vlan_ethtool_get_drvinfo,
@@ -769,6 +775,7 @@ static const struct net_device_ops vlan_netdev_ops = {
 #endif
 	.ndo_fix_features	= vlan_dev_fix_features,
 	.ndo_get_lock_subclass  = vlan_dev_get_lock_subclass,
+	.ndo_get_iflink		= vlan_dev_get_iflink,
 };
 
 static void vlan_dev_free(struct net_device *dev)
-- 
2.2.2

[PATCH net-next v2 06/12] macvlan: implement ndo_get_iflink

[Nicolas Dichtel]

Don't use dev->iflink anymore.

CC: Patrick McHardy <kaber@trash.net>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 drivers/net/macvlan.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c
index b5e3320ca506..b227a13f6473 100644
--- a/drivers/net/macvlan.c
+++ b/drivers/net/macvlan.c
@@ -786,7 +786,6 @@ static int macvlan_init(struct net_device *dev)
 	dev->hw_features	|= NETIF_F_LRO;
 	dev->vlan_features	= lowerdev->vlan_features & MACVLAN_FEATURES;
 	dev->gso_max_size	= lowerdev->gso_max_size;
-	dev->iflink		= lowerdev->ifindex;
 	dev->hard_header_len	= lowerdev->hard_header_len;
 
 	macvlan_set_lockdep_class(dev);
@@ -995,6 +994,13 @@ static void macvlan_dev_netpoll_cleanup(struct net_device *dev)
 }
 #endif	/* CONFIG_NET_POLL_CONTROLLER */
 
+static int macvlan_dev_get_iflink(const struct net_device *dev)
+{
+	struct macvlan_dev *vlan = netdev_priv(dev);
+
+	return vlan->lowerdev->ifindex;
+}
+
 static const struct ethtool_ops macvlan_ethtool_ops = {
 	.get_link		= ethtool_op_get_link,
 	.get_settings		= macvlan_ethtool_get_settings,
@@ -1025,6 +1031,7 @@ static const struct net_device_ops macvlan_netdev_ops = {
 	.ndo_netpoll_setup	= macvlan_dev_netpoll_setup,
 	.ndo_netpoll_cleanup	= macvlan_dev_netpoll_cleanup,
 #endif
+	.ndo_get_iflink		= macvlan_dev_get_iflink,
 };
 
 void macvlan_common_setup(struct net_device *dev)
-- 
2.2.2

[PATCH net-next v2 07/12] ipvlan: implement ndo_get_iflink

[Nicolas Dichtel]

Don't use dev->iflink anymore.

CC: Mahesh Bandewar <maheshb@google.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 drivers/net/ipvlan/ipvlan_main.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ipvlan/ipvlan_main.c b/drivers/net/ipvlan/ipvlan_main.c
index 2950c3780230..1701ede2df89 100644
--- a/drivers/net/ipvlan/ipvlan_main.c
+++ b/drivers/net/ipvlan/ipvlan_main.c
@@ -114,7 +114,6 @@ static int ipvlan_init(struct net_device *dev)
 	dev->features = phy_dev->features & IPVLAN_FEATURES;
 	dev->features |= NETIF_F_LLTX;
 	dev->gso_max_size = phy_dev->gso_max_size;
-	dev->iflink = phy_dev->ifindex;
 	dev->hard_header_len = phy_dev->hard_header_len;
 
 	ipvlan_set_lockdep_class(dev);
@@ -305,6 +304,13 @@ static int ipvlan_vlan_rx_kill_vid(struct net_device *dev, __be16 proto,
 	return 0;
 }
 
+static int ipvlan_get_iflink(const struct net_device *dev)
+{
+	struct ipvl_dev *ipvlan = netdev_priv(dev);
+
+	return ipvlan->phy_dev->ifindex;
+}
+
 static const struct net_device_ops ipvlan_netdev_ops = {
 	.ndo_init		= ipvlan_init,
 	.ndo_uninit		= ipvlan_uninit,
@@ -317,6 +323,7 @@ static const struct net_device_ops ipvlan_netdev_ops = {
 	.ndo_get_stats64	= ipvlan_get_stats64,
 	.ndo_vlan_rx_add_vid	= ipvlan_vlan_rx_add_vid,
 	.ndo_vlan_rx_kill_vid	= ipvlan_vlan_rx_kill_vid,
+	.ndo_get_iflink		= ipvlan_get_iflink,
 };
 
 static int ipvlan_hard_header(struct sk_buff *skb, struct net_device *dev,
-- 
2.2.2

[PATCH net-next v2 08/12] infiniband/ipoib: implement ndo_get_iflink

[Nicolas Dichtel]

Don't use dev->iflink anymore.

CC: Roland Dreier <roland@kernel.org>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 drivers/infiniband/ulp/ipoib/ipoib_main.c | 8 ++++++++
 drivers/infiniband/ulp/ipoib/ipoib_vlan.c | 1 -
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 58b5aa3b6f2d..657b89b1d291 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -842,6 +842,13 @@ static void ipoib_set_mcast_list(struct net_device *dev)
 	queue_work(ipoib_workqueue, &priv->restart_task);
 }
 
+static int ipoib_get_iflink(const struct net_device *dev)
+{
+	struct ipoib_dev_priv *priv = netdev_priv(dev);
+
+	return priv->parent->ifindex;
+}
+
 static u32 ipoib_addr_hash(struct ipoib_neigh_hash *htbl, u8 *daddr)
 {
 	/*
@@ -1341,6 +1348,7 @@ static const struct net_device_ops ipoib_netdev_ops = {
 	.ndo_start_xmit	 	 = ipoib_start_xmit,
 	.ndo_tx_timeout		 = ipoib_timeout,
 	.ndo_set_rx_mode	 = ipoib_set_mcast_list,
+	.ndo_get_iflink		 = ipoib_get_iflink,
 };
 
 void ipoib_setup(struct net_device *dev)
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
index 9fad7b5ac8b9..4dd1313056a4 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
@@ -102,7 +102,6 @@ int __ipoib_vlan_add(struct ipoib_dev_priv *ppriv, struct ipoib_dev_priv *priv,
 	}
 
 	priv->child_type  = type;
-	priv->dev->iflink = ppriv->dev->ifindex;
 	list_add_tail(&priv->list, &ppriv->child_intfs);
 
 	return 0;
-- 
2.2.2

[PATCH net-next v2 09/12] dsa: implement ndo_get_iflink

[Nicolas Dichtel]

Don't use dev->iflink anymore.

CC: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 net/dsa/slave.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/net/dsa/slave.c b/net/dsa/slave.c
index 3597724ec3d8..827cda560a55 100644
--- a/net/dsa/slave.c
+++ b/net/dsa/slave.c
@@ -55,13 +55,11 @@ void dsa_slave_mii_bus_init(struct dsa_switch *ds)
 
 
 /* slave device handling ****************************************************/
-static int dsa_slave_init(struct net_device *dev)
+static int dsa_slave_get_iflink(const struct net_device *dev)
 {
 	struct dsa_slave_priv *p = netdev_priv(dev);
 
-	dev->iflink = p->parent->dst->master_netdev->ifindex;
-
-	return 0;
+	return p->parent->dst->master_netdev->ifindex;
 }
 
 static inline bool dsa_port_is_bridged(struct dsa_slave_priv *p)
@@ -664,7 +662,6 @@ static const struct ethtool_ops dsa_slave_ethtool_ops = {
 };
 
 static const struct net_device_ops dsa_slave_netdev_ops = {
-	.ndo_init		= dsa_slave_init,
 	.ndo_open	 	= dsa_slave_open,
 	.ndo_stop		= dsa_slave_close,
 	.ndo_start_xmit		= dsa_slave_xmit,
@@ -675,6 +672,7 @@ static const struct net_device_ops dsa_slave_netdev_ops = {
 	.ndo_fdb_del		= dsa_slave_fdb_del,
 	.ndo_fdb_dump		= dsa_slave_fdb_dump,
 	.ndo_do_ioctl		= dsa_slave_ioctl,
+	.ndo_get_iflink		= dsa_slave_get_iflink,
 };
 
 static const struct swdev_ops dsa_slave_swdev_ops = {
-- 
2.2.2

[PATCH net-next v2 10/12] net: remove iflink field from struct net_device

[Nicolas Dichtel]

Now that all users of iflink have the ndo_get_iflink handler available, it's
possible to remove this field.

By default, dev_get_iflink() returns the ifindex of the interface.

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 include/linux/netdevice.h |  3 +--
 net/core/dev.c            | 13 ++-----------
 2 files changed, 3 insertions(+), 13 deletions(-)

diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 788eb7a622ad..846a1f5bc9db 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -1538,7 +1538,7 @@ struct net_device {
 	netdev_features_t	mpls_features;
 
 	int			ifindex;
-	int			iflink;
+	int			group;
 
 	struct net_device_stats	stats;
 
@@ -1741,7 +1741,6 @@ struct net_device {
 #endif
 	struct phy_device *phydev;
 	struct lock_class_key *qdisc_tx_busylock;
-	int group;
 	struct pm_qos_request	pm_qos_req;
 };
 #define to_net_dev(d) container_of(d, struct net_device, dev)
diff --git a/net/core/dev.c b/net/core/dev.c
index 77172d085760..3be107e0bc93 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -672,7 +672,7 @@ int dev_get_iflink(const struct net_device *dev)
 	if (dev->netdev_ops && dev->netdev_ops->ndo_get_iflink)
 		return dev->netdev_ops->ndo_get_iflink(dev);
 
-	return dev->iflink;
+	return dev->ifindex;
 }
 EXPORT_SYMBOL(dev_get_iflink);
 
@@ -6331,8 +6331,6 @@ int register_netdevice(struct net_device *dev)
 	spin_lock_init(&dev->addr_list_lock);
 	netdev_set_addr_lockdep_class(dev);
 
-	dev->iflink = -1;
-
 	ret = dev_get_valid_name(net, dev, dev->name);
 	if (ret < 0)
 		goto out;
@@ -6362,9 +6360,6 @@ int register_netdevice(struct net_device *dev)
 	else if (__dev_get_by_index(net, dev->ifindex))
 		goto err_uninit;
 
-	if (dev_get_iflink(dev) == -1)
-		dev->iflink = dev->ifindex;
-
 	/* Transfer changeable features to wanted_features and enable
 	 * software offloads (GSO and GRO).
 	 */
@@ -7077,12 +7072,8 @@ int dev_change_net_namespace(struct net_device *dev, struct net *net, const char
 	dev_net_set(dev, net);
 
 	/* If there is an ifindex conflict assign a new one */
-	if (__dev_get_by_index(net, dev->ifindex)) {
-		int iflink = (dev_get_iflink(dev) == dev->ifindex);
+	if (__dev_get_by_index(net, dev->ifindex))
 		dev->ifindex = dev_new_index(net);
-		if (iflink)
-			dev->iflink = dev->ifindex;
-	}
 
 	/* Send a netdev-add uevent to the new namespace */
 	kobject_uevent(&dev->dev.kobj, KOBJ_ADD);
-- 
2.2.2

[PATCH net-next v2 11/12] dev: set iflink to 0 for virtual interfaces

[Nicolas Dichtel]

Virtual interfaces are supposed to set an iflink value != of their ifindex.
It was not the case for some of them, like vxlan, bond or bridge.
Let's set iflink to 0 when dev->rtnl_link_ops is set.

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 net/core/dev.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/net/core/dev.c b/net/core/dev.c
index 3be107e0bc93..26622d614f81 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -672,6 +672,10 @@ int dev_get_iflink(const struct net_device *dev)
 	if (dev->netdev_ops && dev->netdev_ops->ndo_get_iflink)
 		return dev->netdev_ops->ndo_get_iflink(dev);
 
+	/* If dev->rtnl_link_ops is set, it's a virtual interface. */
+	if (dev->rtnl_link_ops)
+		return 0;
+
 	return dev->ifindex;
 }
 EXPORT_SYMBOL(dev_get_iflink);
-- 
2.2.2

[PATCH net-next v2 12/12] veth: set iflink to the peer veth

[Nicolas Dichtel]

Now that the peer netns is advertised in rtnl messages, we can set this property
so that IFLA_LINK will advertise the peer ifindex. It allows the userland to get
the full veth configuration.

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
---
 drivers/net/veth.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/drivers/net/veth.c b/drivers/net/veth.c
index 4cca36ebc4fb..c8186ffda1a3 100644
--- a/drivers/net/veth.c
+++ b/drivers/net/veth.c
@@ -263,6 +263,20 @@ static void veth_poll_controller(struct net_device *dev)
 }
 #endif	/* CONFIG_NET_POLL_CONTROLLER */
 
+static int veth_get_iflink(const struct net_device *dev)
+{
+	struct veth_priv *priv = netdev_priv(dev);
+	struct net_device *peer;
+	int iflink;
+
+	rcu_read_lock();
+	peer = rcu_dereference(priv->peer);
+	iflink = peer ? peer->ifindex : 0;
+	rcu_read_unlock();
+
+	return iflink;
+}
+
 static const struct net_device_ops veth_netdev_ops = {
 	.ndo_init            = veth_dev_init,
 	.ndo_open            = veth_open,
@@ -275,6 +289,7 @@ static const struct net_device_ops veth_netdev_ops = {
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	.ndo_poll_controller	= veth_poll_controller,
 #endif
+	.ndo_get_iflink		= veth_get_iflink,
 };
 
 #define VETH_FEATURES (NETIF_F_SG | NETIF_F_FRAGLIST | NETIF_F_ALL_TSO |    \
-- 
2.2.2

Re: [PATCH net-next v2 0/12] Remove iflink field from the net_device structure

[David Miller]

From: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Date: Thu,  2 Apr 2015 17:06:59 +0200

> The first goal of this series was to advertise the veth peer via the IFLA_LINK
> attribute, but iflink was not ready for network namespaces.
> 
> The iflink of an interface should be set to its ifindex for a physical interface
> and to another value (0 if not relevant) for a virtual interface.
> This was not the case for some interfaces, like vxlan, bond, or bridge for
> example.
> There is also a risk, if the targeted interface moves to another netns, that the
> ifindex changes without updating corresponding iflink fields (eg. vlan).
> 
> Moving the management of this property into virtual interface drivers allows to
> better handle this last case because most of virtual interface drivers have a
> pointer to the link netdevice.
> Anyway, dev->iflink value was always a copy of some internal data of the virtual
> interface driver, thus let's use these internal data directly.
> 
> So, this series removes the iflink field and let the drivers manage it.
> Only the last patch was present in the v1, but I fully rework it.

I like this a lot, series applied, thanks Nicolas.