[net-next,2/4] vxlan: support setting IPv6 flow label

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>
---
 drivers/net/vxlan.c          | 26 +++++++++++++++++++++-----
 include/net/vxlan.h          |  1 +
 include/uapi/linux/if_link.h |  1 +
 3 files changed, 23 insertions(+), 5 deletions(-)

On Tue, Mar 8, 2016 at 6:00 PM, Daniel Borkmann <daniel@iogearbox.net> wrote:
> 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>
> ---
>  drivers/net/vxlan.c          | 26 +++++++++++++++++++++-----
>  include/net/vxlan.h          |  1 +
>  include/uapi/linux/if_link.h |  1 +
>  3 files changed, 23 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
> index 8bdcd5e..8eda76f 100644
> --- a/drivers/net/vxlan.c
> +++ b/drivers/net/vxlan.c
> @@ -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,
> diff --git a/include/net/vxlan.h b/include/net/vxlan.h
> index 6eda4ed..a763c96 100644
> --- a/include/net/vxlan.h
> +++ b/include/net/vxlan.h
> @@ -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;
> diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h
> index d452cea..6bebc97 100644
> --- a/include/uapi/linux/if_link.h
> +++ b/include/uapi/linux/if_link.h
> @@ -456,6 +456,7 @@ enum {
>         IFLA_VXLAN_GBP,
>         IFLA_VXLAN_REMCSUM_NOPARTIAL,
>         IFLA_VXLAN_COLLECT_METADATA,
> +       IFLA_VXLAN_LABEL,

Should this interface align with ip_gre that uses IFLA_GRE_FLOWINFO to
set both traffic class and flow label in one shot?

>         __IFLA_VXLAN_MAX
>  };
>  #define IFLA_VXLAN_MAX (__IFLA_VXLAN_MAX - 1)
> --
> 1.9.3
>

On 03/09/2016 04:29 AM, Tom Herbert wrote:
[...]
>> diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h
>> index d452cea..6bebc97 100644
>> --- a/include/uapi/linux/if_link.h
>> +++ b/include/uapi/linux/if_link.h
>> @@ -456,6 +456,7 @@ enum {
>>          IFLA_VXLAN_GBP,
>>          IFLA_VXLAN_REMCSUM_NOPARTIAL,
>>          IFLA_VXLAN_COLLECT_METADATA,
>> +       IFLA_VXLAN_LABEL,
>
> Should this interface align with ip_gre that uses IFLA_GRE_FLOWINFO to
> set both traffic class and flow label in one shot?

For vxlan and geneve by the way, there's already IFLA_VXLAN_TOS and
IFLA_GENEVE_TOS attribute that configure the tos for IPv4 or traffic
class for IPv6 case on the device, and have special handling in the
xmit path that they can inherit the tos/tc when configured as such on
the device in non collect metadata case. Seems like grev6 is handled
differently and also still without tunnel key support, if I see this
correctly. Agree with you that some more consolidation would be nice.

>>          __IFLA_VXLAN_MAX
>>   };
>>   #define IFLA_VXLAN_MAX (__IFLA_VXLAN_MAX - 1)
>> --
>> 1.9.3