Eric Woudstra <[email protected]> wrote:
> This adds the capability to evaluate 802.1ad, QinQ, PPPoE and PPPoE-in-Q
> packets in the bridge filter chain.
>
> Signed-off-by: Eric Woudstra <[email protected]>
> ---
> net/netfilter/nft_chain_filter.c | 52 +++++++++++++++++++++++++++++++-
> 1 file changed, 51 insertions(+), 1 deletion(-)
>
> diff --git a/net/netfilter/nft_chain_filter.c
> b/net/netfilter/nft_chain_filter.c
> index 19a553550c76..8445ddfb9cea 100644
> --- a/net/netfilter/nft_chain_filter.c
> +++ b/net/netfilter/nft_chain_filter.c
> @@ -232,11 +232,55 @@ nft_do_chain_bridge(void *priv,
> struct sk_buff *skb,
> const struct nf_hook_state *state)
> {
> + __be16 outer_proto, proto = 0;
> struct nft_pktinfo pkt;
> + int ret, offset = 0;
>
> nft_set_pktinfo(&pkt, skb, state);
>
> switch (eth_hdr(skb)->h_proto) {
> + case htons(ETH_P_PPP_SES): {
> + struct ppp_hdr {
> + struct pppoe_hdr hdr;
> + __be16 proto;
> + } *ph;
> +
> + if (!pskb_may_pull(skb, PPPOE_SES_HLEN))
> + break;
> + offset = PPPOE_SES_HLEN;
> + outer_proto = skb->protocol;
> + ph = (struct ppp_hdr *)(skb->data);
> + switch (ph->proto) {
> + case htons(PPP_IP):
> + proto = htons(ETH_P_IP);
> + break;
> + case htons(PPP_IPV6):
> + proto = htons(ETH_P_IPV6);
> + break;
> + }
What if ph->proto is neither ipv4 nor ipv6? I don't think
we should clobber skb->protocol or set a new network header in that
case.
> + ret = nft_do_chain(&pkt, priv);
> +
> + if (offset) {
> + skb_reset_network_header(skb);
if ret == NF_STOLEN, skb has already been free'd or is queued
elsewhere and must not be modified anymore.
I would restrict this to ret == NF_ACCEPT.
