vxlan: support setting IPv6 flow label
authorDaniel Borkmann <daniel@iogearbox.net>
Wed, 9 Mar 2016 02:00:03 +0000 (03:00 +0100)
committerDavid S. Miller <davem@davemloft.net>
Fri, 11 Mar 2016 20:14:26 +0000 (15:14 -0500)
This work adds support for setting the IPv6 flow label for vxlan per
device and through collect metadata (ip_tunnel_key) frontends. The
vxlan dst cache does not need any special considerations here, for
the cases where caches can be used, the label is static per cache.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/vxlan.c
include/net/vxlan.h
include/uapi/linux/if_link.h

index 8bdcd5ea84240e0c1bd69080e3170fddf7312fae..8eda76f9e474ddaa29e2db68fcae9f1e16f1fd39 100644 (file)
@@ -1790,6 +1790,7 @@ static struct rtable *vxlan_get_route(struct vxlan_dev *vxlan,
 #if IS_ENABLED(CONFIG_IPV6)
 static struct dst_entry *vxlan6_get_route(struct vxlan_dev *vxlan,
                                          struct sk_buff *skb, int oif, u8 tos,
+                                         __be32 label,
                                          const struct in6_addr *daddr,
                                          struct in6_addr *saddr,
                                          struct dst_cache *dst_cache,
@@ -1813,6 +1814,7 @@ static struct dst_entry *vxlan6_get_route(struct vxlan_dev *vxlan,
        fl6.flowi6_tos = RT_TOS(tos);
        fl6.daddr = *daddr;
        fl6.saddr = vxlan->cfg.saddr.sin6.sin6_addr;
+       fl6.flowlabel = label;
        fl6.flowi6_mark = skb->mark;
        fl6.flowi6_proto = IPPROTO_UDP;
 
@@ -1888,7 +1890,7 @@ static void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
        struct vxlan_metadata _md;
        struct vxlan_metadata *md = &_md;
        __be16 src_port = 0, dst_port;
-       __be32 vni;
+       __be32 vni, label;
        __be16 df = 0;
        __u8 tos, ttl;
        int err;
@@ -1939,12 +1941,14 @@ static void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
        if (tos == 1)
                tos = ip_tunnel_get_dsfield(old_iph, skb);
 
+       label = vxlan->cfg.label;
        src_port = udp_flow_src_port(dev_net(dev), skb, vxlan->cfg.port_min,
                                     vxlan->cfg.port_max, true);
 
        if (info) {
                ttl = info->key.ttl;
                tos = info->key.tos;
+               label = info->key.label;
                udp_sum = !!(info->key.tun_flags & TUNNEL_CSUM);
 
                if (info->options_len)
@@ -2020,7 +2024,7 @@ static void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 
                ndst = vxlan6_get_route(vxlan, skb,
                                        rdst ? rdst->remote_ifindex : 0, tos,
-                                       &dst->sin6.sin6_addr, &saddr,
+                                       label, &dst->sin6.sin6_addr, &saddr,
                                        dst_cache, info);
                if (IS_ERR(ndst)) {
                        netdev_dbg(dev, "no route to %pI6\n",
@@ -2066,8 +2070,8 @@ static void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
                        return;
                }
                udp_tunnel6_xmit_skb(ndst, sk, skb, dev,
-                                    &saddr, &dst->sin6.sin6_addr, tos, ttl, 0,
-                                    src_port, dst_port, !udp_sum);
+                                    &saddr, &dst->sin6.sin6_addr, tos, ttl,
+                                    label, src_port, dst_port, !udp_sum);
 #endif
        }
 
@@ -2390,7 +2394,7 @@ static int vxlan_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
                if (!vxlan->vn6_sock)
                        return -EINVAL;
                ndst = vxlan6_get_route(vxlan, skb, 0, info->key.tos,
-                                       &info->key.u.ipv6.dst,
+                                       info->key.label, &info->key.u.ipv6.dst,
                                        &info->key.u.ipv6.src, NULL, info);
                if (IS_ERR(ndst))
                        return PTR_ERR(ndst);
@@ -2505,6 +2509,7 @@ static const struct nla_policy vxlan_policy[IFLA_VXLAN_MAX + 1] = {
        [IFLA_VXLAN_LOCAL6]     = { .len = sizeof(struct in6_addr) },
        [IFLA_VXLAN_TOS]        = { .type = NLA_U8 },
        [IFLA_VXLAN_TTL]        = { .type = NLA_U8 },
+       [IFLA_VXLAN_LABEL]      = { .type = NLA_U32 },
        [IFLA_VXLAN_LEARNING]   = { .type = NLA_U8 },
        [IFLA_VXLAN_AGEING]     = { .type = NLA_U32 },
        [IFLA_VXLAN_LIMIT]      = { .type = NLA_U32 },
@@ -2739,6 +2744,11 @@ static int vxlan_dev_configure(struct net *src_net, struct net_device *dev,
                vxlan->flags |= VXLAN_F_IPV6;
        }
 
+       if (conf->label && !use_ipv6) {
+               pr_info("label only supported in use with IPv6\n");
+               return -EINVAL;
+       }
+
        if (conf->remote_ifindex) {
                lowerdev = __dev_get_by_index(src_net, conf->remote_ifindex);
                dst->remote_ifindex = conf->remote_ifindex;
@@ -2887,6 +2897,10 @@ static int vxlan_newlink(struct net *src_net, struct net_device *dev,
        if (data[IFLA_VXLAN_TTL])
                conf.ttl = nla_get_u8(data[IFLA_VXLAN_TTL]);
 
+       if (data[IFLA_VXLAN_LABEL])
+               conf.label = nla_get_be32(data[IFLA_VXLAN_LABEL]) &
+                            IPV6_FLOWLABEL_MASK;
+
        if (!data[IFLA_VXLAN_LEARNING] || nla_get_u8(data[IFLA_VXLAN_LEARNING]))
                conf.flags |= VXLAN_F_LEARN;
 
@@ -2990,6 +3004,7 @@ static size_t vxlan_get_size(const struct net_device *dev)
                nla_total_size(sizeof(struct in6_addr)) + /* IFLA_VXLAN_LOCAL{6} */
                nla_total_size(sizeof(__u8)) +  /* IFLA_VXLAN_TTL */
                nla_total_size(sizeof(__u8)) +  /* IFLA_VXLAN_TOS */
+               nla_total_size(sizeof(__be32)) + /* IFLA_VXLAN_LABEL */
                nla_total_size(sizeof(__u8)) +  /* IFLA_VXLAN_LEARNING */
                nla_total_size(sizeof(__u8)) +  /* IFLA_VXLAN_PROXY */
                nla_total_size(sizeof(__u8)) +  /* IFLA_VXLAN_RSC */
@@ -3053,6 +3068,7 @@ static int vxlan_fill_info(struct sk_buff *skb, const struct net_device *dev)
 
        if (nla_put_u8(skb, IFLA_VXLAN_TTL, vxlan->cfg.ttl) ||
            nla_put_u8(skb, IFLA_VXLAN_TOS, vxlan->cfg.tos) ||
+           nla_put_be32(skb, IFLA_VXLAN_LABEL, vxlan->cfg.label) ||
            nla_put_u8(skb, IFLA_VXLAN_LEARNING,
                        !!(vxlan->flags & VXLAN_F_LEARN)) ||
            nla_put_u8(skb, IFLA_VXLAN_PROXY,
index 6eda4ed4d78bc03cf99eb551ebebfc59698238d2..a763c96ecde403bd4cc8ddd9f6136685c281be2c 100644 (file)
@@ -162,6 +162,7 @@ struct vxlan_config {
        u16                     port_max;
        u8                      tos;
        u8                      ttl;
+       __be32                  label;
        u32                     flags;
        unsigned long           age_interval;
        unsigned int            addrmax;
index d452cea5902039e2abd15cbd344a75fd879a6228..6bebc975031d6c6698b23705bc0205381950d372 100644 (file)
@@ -456,6 +456,7 @@ enum {
        IFLA_VXLAN_GBP,
        IFLA_VXLAN_REMCSUM_NOPARTIAL,
        IFLA_VXLAN_COLLECT_METADATA,
+       IFLA_VXLAN_LABEL,
        __IFLA_VXLAN_MAX
 };
 #define IFLA_VXLAN_MAX (__IFLA_VXLAN_MAX - 1)