ipv6: Set skb->protocol properly for local output
authorEli Cooper <elicooper@gmx.com>
Thu, 1 Dec 2016 02:05:11 +0000 (10:05 +0800)
committerBen Hutchings <ben@decadent.org.uk>
Thu, 23 Feb 2017 03:54:41 +0000 (03:54 +0000)
commit b4e479a96fc398ccf83bb1cffb4ffef8631beaf1 upstream.

When xfrm is applied to TSO/GSO packets, it follows this path:

    xfrm_output() -> xfrm_output_gso() -> skb_gso_segment()

where skb_gso_segment() relies on skb->protocol to function properly.

This patch sets skb->protocol to ETH_P_IPV6 before dst_output() is called,
fixing a bug where GSO packets sent through an ipip6 tunnel are dropped
when xfrm is involved.

Signed-off-by: Eli Cooper <elicooper@gmx.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
net/ipv6/output_core.c

index 1d4156ddf355af7249fac2818791d6b881e50b71..1e987e68e510f47c046100e177f4b3cb47fcb132 100644 (file)
@@ -114,6 +114,8 @@ int __ip6_local_out(struct sk_buff *skb)
        ipv6_hdr(skb)->payload_len = htons(len);
        IP6CB(skb)->nhoff = offsetof(struct ipv6hdr, nexthdr);
 
+       skb->protocol = htons(ETH_P_IPV6);
+
        return nf_hook(NFPROTO_IPV6, NF_INET_LOCAL_OUT, skb, NULL,
                       skb_dst(skb)->dev, dst_output);
 }