netfilter: nf_conntrack_icmpv6: find conntrack related to ICMPv6 redirect packet
Subsystems (2)
PATCH 1/1 netfilter: nf_conntrack_icmpv6: find conntrack related to ICMPv6 redirect packet
--- net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c | 77 ++++++++++++++++++++++++-- 1 file changed, 72 insertions(+), 5 deletions(-)
net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c +72 -5
@@ -144,7 +144,7 @@ static bool icmpv6_new(struct nf_conn *ct, const struct sk_buff *skb,static int icmpv6_error_message (struct net *net,struct nf_conn *tmpl,struct sk_buff *skb, -unsigned int icmp6off, +unsigned int inneripv6off,enum ip_conntrack_info *ctinfo,unsigned int hooknum) {
@@ -157,9 +157,7 @@ icmpv6_error_message(struct net *net, struct nf_conn *tmpl,/* Are they talking about one of our connections? */ if (!nf_ct_get_tuplepr (skb, -skb_network_offset (skb) - +sizeof (struct ipv6hdr ) - +sizeof (struct icmp6hdr ), + inneripv6off,PF_INET6 , net, &origtuple)) {pr_debug ("icmpv6_error: Can't get tuple\n" );return -NF_ACCEPT ;
@@ -227,9 +225,78 @@ icmpv6_error(struct net *net, struct nf_conn *tmpl,nf_conntrack_get (skb->nfct);return NF_ACCEPT ; } + dataoff +=sizeof (struct icmp6hdr );/* is not error message ? */ -if (icmp6h->icmp6_type >=128 ) +if (icmp6h->icmp6_type ==NDISC_REDIRECT ) { +const struct in6_addr *dst; +struct in6_addr _dst; +const struct nd_opt_hdr *opt; +struct nd_opt_hdr _opt; +const struct ipv6hdr *iph; +struct ipv6hdr _iph; + +/* skip target address */ + dataoff +=sizeof (_dst); + +/* read destination address */ + dst =skb_header_pointer (skb, dataoff,sizeof (_dst), &_dst); +if (dst ==NULL ) { +if (LOG_INVALID (net,IPPROTO_ICMPV6 )) +nf_log_packet (net,PF_INET6 ,0 , skb,NULL ,NULL ,NULL , +"nf_ct_icmpv6: short redirect packet " ); +return -NF_ACCEPT ; + } +if (ipv6_addr_is_multicast (dst)) { +if (LOG_INVALID (net,IPPROTO_ICMPV6 )) +nf_log_packet (net,PF_INET6 ,0 , skb,NULL ,NULL ,NULL , +"nf_ct_icmpv6: redirect destination address is multicast " ); +return -NF_ACCEPT ; + } + dataoff +=sizeof (_dst); + +/* find redirected header */ +while (1 ) { + opt =skb_header_pointer (skb, dataoff,sizeof (_opt), &_opt); +if (opt ==NULL ) { +if (LOG_INVALID (net,IPPROTO_ICMPV6 )) +nf_log_packet (net,PF_INET6 ,0 , skb,NULL ,NULL ,NULL , +"nf_ct_icmpv6: invalid redirect option " ); +return -NF_ACCEPT ; + } +if (opt->nd_opt_len ==0 ) { +if (LOG_INVALID (net,IPPROTO_ICMPV6 )) +nf_log_packet (net,PF_INET6 ,0 , skb,NULL ,NULL ,NULL , +"nf_ct_icmpv6: invalid redirect option length " ); +return -NF_ACCEPT ; + } + +if (opt->nd_opt_type ==ND_OPT_REDIRECT_HDR ) { + dataoff +=8 ; +break ; + } + + dataoff += opt->nd_opt_len <<3 ; + } + +/* read redirect header */ + iph =skb_header_pointer (skb, dataoff,sizeof (_iph), &_iph); +if (iph ==NULL ) { +if (LOG_INVALID (net,IPPROTO_ICMPV6 )) +nf_log_packet (net,PF_INET6 ,0 , skb,NULL ,NULL ,NULL , +"nf_ct_icmpv6: short redirect header " ); +return -NF_ACCEPT ; + } + +/* validate destination address */ +if (!ipv6_addr_equal (&iph->daddr, dst)) { +if (LOG_INVALID (net,IPPROTO_ICMPV6 )) +nf_log_packet (net,PF_INET6 ,0 , skb,NULL ,NULL ,NULL , +"nf_ct_icmpv6: redirect destination address not matching destination address of redirect header " ); +return -NF_ACCEPT ; + } + } +else if (icmp6h->icmp6_type >=128 )return NF_ACCEPT ;return icmpv6_error_message (net, tmpl, skb, dataoff, ctinfo, hooknum);
Signature · 1 line
1.7.12.4
Pablo Neira Ayuso
maintainerI'd suggest you submit this patch with a proper description.
Thanks.
Quoted 3 lines · Pablo Neira Ayuso
I'd suggest you submit this patch with a proper description.
Thanks.
This patch will modify the conntrack state created by ICMPv6 redirect packets from INVALID (as it is implemented now, skb->nfct remains NULL) to RELATED (like in ICMPv6 errors case). In IPv4 case, ICMP redirects are treated the same way as ICMP errors, so there is no issue. Probably ICMPv6 redirects were not handled because their parsing is not as straightforward as ICMPv6 errors.
I tested it on an older version of kernel, but since nf_conntrack_proto_icmpv6.c remained basically the same, I think the issue would be reproducible even on latest version of kernel.
Cheers,
Alin