netfilter: nf_flow_table: move ipv6 offload hook code to nf_flow_table
authorFelix Fietkau <nbd@nbd.name>
Mon, 26 Feb 2018 09:15:15 +0000 (10:15 +0100)
committerPablo Neira Ayuso <pablo@netfilter.org>
Tue, 24 Apr 2018 08:28:18 +0000 (10:28 +0200)
Useful as preparation for adding iptables support for offload.

Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
net/ipv6/netfilter/nf_flow_table_ipv6.c
net/netfilter/nf_flow_table_ip.c

index 0e6328490142b76267872c2c6efa03b4adb66679..f1804ce8d56165338d39d74699bdb891b3d5c92a 100644 (file)
@@ -3,240 +3,8 @@
 #include <linux/module.h>
 #include <linux/netfilter.h>
 #include <linux/rhashtable.h>
-#include <linux/ipv6.h>
-#include <linux/netdevice.h>
-#include <net/ipv6.h>
-#include <net/ip6_route.h>
-#include <net/neighbour.h>
 #include <net/netfilter/nf_flow_table.h>
 #include <net/netfilter/nf_tables.h>
-/* For layer 4 checksum field offset. */
-#include <linux/tcp.h>
-#include <linux/udp.h>
-
-static int nf_flow_nat_ipv6_tcp(struct sk_buff *skb, unsigned int thoff,
-                               struct in6_addr *addr,
-                               struct in6_addr *new_addr)
-{
-       struct tcphdr *tcph;
-
-       if (!pskb_may_pull(skb, thoff + sizeof(*tcph)) ||
-           skb_try_make_writable(skb, thoff + sizeof(*tcph)))
-               return -1;
-
-       tcph = (void *)(skb_network_header(skb) + thoff);
-       inet_proto_csum_replace16(&tcph->check, skb, addr->s6_addr32,
-                                 new_addr->s6_addr32, true);
-
-       return 0;
-}
-
-static int nf_flow_nat_ipv6_udp(struct sk_buff *skb, unsigned int thoff,
-                               struct in6_addr *addr,
-                               struct in6_addr *new_addr)
-{
-       struct udphdr *udph;
-
-       if (!pskb_may_pull(skb, thoff + sizeof(*udph)) ||
-           skb_try_make_writable(skb, thoff + sizeof(*udph)))
-               return -1;
-
-       udph = (void *)(skb_network_header(skb) + thoff);
-       if (udph->check || skb->ip_summed == CHECKSUM_PARTIAL) {
-               inet_proto_csum_replace16(&udph->check, skb, addr->s6_addr32,
-                                         new_addr->s6_addr32, true);
-               if (!udph->check)
-                       udph->check = CSUM_MANGLED_0;
-       }
-
-       return 0;
-}
-
-static int nf_flow_nat_ipv6_l4proto(struct sk_buff *skb, struct ipv6hdr *ip6h,
-                                   unsigned int thoff, struct in6_addr *addr,
-                                   struct in6_addr *new_addr)
-{
-       switch (ip6h->nexthdr) {
-       case IPPROTO_TCP:
-               if (nf_flow_nat_ipv6_tcp(skb, thoff, addr, new_addr) < 0)
-                       return NF_DROP;
-               break;
-       case IPPROTO_UDP:
-               if (nf_flow_nat_ipv6_udp(skb, thoff, addr, new_addr) < 0)
-                       return NF_DROP;
-               break;
-       }
-
-       return 0;
-}
-
-static int nf_flow_snat_ipv6(const struct flow_offload *flow,
-                            struct sk_buff *skb, struct ipv6hdr *ip6h,
-                            unsigned int thoff,
-                            enum flow_offload_tuple_dir dir)
-{
-       struct in6_addr addr, new_addr;
-
-       switch (dir) {
-       case FLOW_OFFLOAD_DIR_ORIGINAL:
-               addr = ip6h->saddr;
-               new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_v6;
-               ip6h->saddr = new_addr;
-               break;
-       case FLOW_OFFLOAD_DIR_REPLY:
-               addr = ip6h->daddr;
-               new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_v6;
-               ip6h->daddr = new_addr;
-               break;
-       default:
-               return -1;
-       }
-
-       return nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
-}
-
-static int nf_flow_dnat_ipv6(const struct flow_offload *flow,
-                            struct sk_buff *skb, struct ipv6hdr *ip6h,
-                            unsigned int thoff,
-                            enum flow_offload_tuple_dir dir)
-{
-       struct in6_addr addr, new_addr;
-
-       switch (dir) {
-       case FLOW_OFFLOAD_DIR_ORIGINAL:
-               addr = ip6h->daddr;
-               new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.src_v6;
-               ip6h->daddr = new_addr;
-               break;
-       case FLOW_OFFLOAD_DIR_REPLY:
-               addr = ip6h->saddr;
-               new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_v6;
-               ip6h->saddr = new_addr;
-               break;
-       default:
-               return -1;
-       }
-
-       return nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
-}
-
-static int nf_flow_nat_ipv6(const struct flow_offload *flow,
-                           struct sk_buff *skb,
-                           enum flow_offload_tuple_dir dir)
-{
-       struct ipv6hdr *ip6h = ipv6_hdr(skb);
-       unsigned int thoff = sizeof(*ip6h);
-
-       if (flow->flags & FLOW_OFFLOAD_SNAT &&
-           (nf_flow_snat_port(flow, skb, thoff, ip6h->nexthdr, dir) < 0 ||
-            nf_flow_snat_ipv6(flow, skb, ip6h, thoff, dir) < 0))
-               return -1;
-       if (flow->flags & FLOW_OFFLOAD_DNAT &&
-           (nf_flow_dnat_port(flow, skb, thoff, ip6h->nexthdr, dir) < 0 ||
-            nf_flow_dnat_ipv6(flow, skb, ip6h, thoff, dir) < 0))
-               return -1;
-
-       return 0;
-}
-
-static int nf_flow_tuple_ipv6(struct sk_buff *skb, const struct net_device *dev,
-                             struct flow_offload_tuple *tuple)
-{
-       struct flow_ports *ports;
-       struct ipv6hdr *ip6h;
-       unsigned int thoff;
-
-       if (!pskb_may_pull(skb, sizeof(*ip6h)))
-               return -1;
-
-       ip6h = ipv6_hdr(skb);
-
-       if (ip6h->nexthdr != IPPROTO_TCP &&
-           ip6h->nexthdr != IPPROTO_UDP)
-               return -1;
-
-       thoff = sizeof(*ip6h);
-       if (!pskb_may_pull(skb, thoff + sizeof(*ports)))
-               return -1;
-
-       ports = (struct flow_ports *)(skb_network_header(skb) + thoff);
-
-       tuple->src_v6           = ip6h->saddr;
-       tuple->dst_v6           = ip6h->daddr;
-       tuple->src_port         = ports->source;
-       tuple->dst_port         = ports->dest;
-       tuple->l3proto          = AF_INET6;
-       tuple->l4proto          = ip6h->nexthdr;
-       tuple->iifidx           = dev->ifindex;
-
-       return 0;
-}
-
-/* Based on ip_exceeds_mtu(). */
-static bool nf_flow_exceeds_mtu(const struct sk_buff *skb, unsigned int mtu)
-{
-       if (skb->len <= mtu)
-               return false;
-
-       if (skb_is_gso(skb) && skb_gso_validate_network_len(skb, mtu))
-               return false;
-
-       return true;
-}
-
-unsigned int
-nf_flow_offload_ipv6_hook(void *priv, struct sk_buff *skb,
-                         const struct nf_hook_state *state)
-{
-       struct flow_offload_tuple_rhash *tuplehash;
-       struct nf_flowtable *flow_table = priv;
-       struct flow_offload_tuple tuple = {};
-       enum flow_offload_tuple_dir dir;
-       struct flow_offload *flow;
-       struct net_device *outdev;
-       struct in6_addr *nexthop;
-       struct ipv6hdr *ip6h;
-       struct rt6_info *rt;
-
-       if (skb->protocol != htons(ETH_P_IPV6))
-               return NF_ACCEPT;
-
-       if (nf_flow_tuple_ipv6(skb, state->in, &tuple) < 0)
-               return NF_ACCEPT;
-
-       tuplehash = flow_offload_lookup(flow_table, &tuple);
-       if (tuplehash == NULL)
-               return NF_ACCEPT;
-
-       outdev = dev_get_by_index_rcu(state->net, tuplehash->tuple.oifidx);
-       if (!outdev)
-               return NF_ACCEPT;
-
-       dir = tuplehash->tuple.dir;
-       flow = container_of(tuplehash, struct flow_offload, tuplehash[dir]);
-       rt = (struct rt6_info *)flow->tuplehash[dir].tuple.dst_cache;
-
-       if (unlikely(nf_flow_exceeds_mtu(skb, flow->tuplehash[dir].tuple.mtu)))
-               return NF_ACCEPT;
-
-       if (skb_try_make_writable(skb, sizeof(*ip6h)))
-               return NF_DROP;
-
-       if (flow->flags & (FLOW_OFFLOAD_SNAT | FLOW_OFFLOAD_DNAT) &&
-           nf_flow_nat_ipv6(flow, skb, dir) < 0)
-               return NF_DROP;
-
-       flow->timeout = (u32)jiffies + NF_FLOW_TIMEOUT;
-       ip6h = ipv6_hdr(skb);
-       ip6h->hop_limit--;
-
-       skb->dev = outdev;
-       nexthop = rt6_nexthop(rt, &flow->tuplehash[!dir].tuple.src_v6);
-       neigh_xmit(NEIGH_ND_TABLE, outdev, nexthop, skb);
-
-       return NF_STOLEN;
-}
-EXPORT_SYMBOL_GPL(nf_flow_offload_ipv6_hook);
 
 static struct nf_flowtable_type flowtable_ipv6 = {
        .family         = NFPROTO_IPV6,
index 103263e0c7c2ed4afa355105c9496b6a969fe693..dc570fb7641dd718cb8d42bab1de31b5c9fc53d0 100644 (file)
@@ -4,8 +4,11 @@
 #include <linux/netfilter.h>
 #include <linux/rhashtable.h>
 #include <linux/ip.h>
+#include <linux/ipv6.h>
 #include <linux/netdevice.h>
 #include <net/ip.h>
+#include <net/ipv6.h>
+#include <net/ip6_route.h>
 #include <net/neighbour.h>
 #include <net/netfilter/nf_flow_table.h>
 /* For layer 4 checksum field offset. */
@@ -242,3 +245,215 @@ nf_flow_offload_ip_hook(void *priv, struct sk_buff *skb,
        return NF_STOLEN;
 }
 EXPORT_SYMBOL_GPL(nf_flow_offload_ip_hook);
+
+static int nf_flow_nat_ipv6_tcp(struct sk_buff *skb, unsigned int thoff,
+                               struct in6_addr *addr,
+                               struct in6_addr *new_addr)
+{
+       struct tcphdr *tcph;
+
+       if (!pskb_may_pull(skb, thoff + sizeof(*tcph)) ||
+           skb_try_make_writable(skb, thoff + sizeof(*tcph)))
+               return -1;
+
+       tcph = (void *)(skb_network_header(skb) + thoff);
+       inet_proto_csum_replace16(&tcph->check, skb, addr->s6_addr32,
+                                 new_addr->s6_addr32, true);
+
+       return 0;
+}
+
+static int nf_flow_nat_ipv6_udp(struct sk_buff *skb, unsigned int thoff,
+                               struct in6_addr *addr,
+                               struct in6_addr *new_addr)
+{
+       struct udphdr *udph;
+
+       if (!pskb_may_pull(skb, thoff + sizeof(*udph)) ||
+           skb_try_make_writable(skb, thoff + sizeof(*udph)))
+               return -1;
+
+       udph = (void *)(skb_network_header(skb) + thoff);
+       if (udph->check || skb->ip_summed == CHECKSUM_PARTIAL) {
+               inet_proto_csum_replace16(&udph->check, skb, addr->s6_addr32,
+                                         new_addr->s6_addr32, true);
+               if (!udph->check)
+                       udph->check = CSUM_MANGLED_0;
+       }
+
+       return 0;
+}
+
+static int nf_flow_nat_ipv6_l4proto(struct sk_buff *skb, struct ipv6hdr *ip6h,
+                                   unsigned int thoff, struct in6_addr *addr,
+                                   struct in6_addr *new_addr)
+{
+       switch (ip6h->nexthdr) {
+       case IPPROTO_TCP:
+               if (nf_flow_nat_ipv6_tcp(skb, thoff, addr, new_addr) < 0)
+                       return NF_DROP;
+               break;
+       case IPPROTO_UDP:
+               if (nf_flow_nat_ipv6_udp(skb, thoff, addr, new_addr) < 0)
+                       return NF_DROP;
+               break;
+       }
+
+       return 0;
+}
+
+static int nf_flow_snat_ipv6(const struct flow_offload *flow,
+                            struct sk_buff *skb, struct ipv6hdr *ip6h,
+                            unsigned int thoff,
+                            enum flow_offload_tuple_dir dir)
+{
+       struct in6_addr addr, new_addr;
+
+       switch (dir) {
+       case FLOW_OFFLOAD_DIR_ORIGINAL:
+               addr = ip6h->saddr;
+               new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_v6;
+               ip6h->saddr = new_addr;
+               break;
+       case FLOW_OFFLOAD_DIR_REPLY:
+               addr = ip6h->daddr;
+               new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_v6;
+               ip6h->daddr = new_addr;
+               break;
+       default:
+               return -1;
+       }
+
+       return nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
+}
+
+static int nf_flow_dnat_ipv6(const struct flow_offload *flow,
+                            struct sk_buff *skb, struct ipv6hdr *ip6h,
+                            unsigned int thoff,
+                            enum flow_offload_tuple_dir dir)
+{
+       struct in6_addr addr, new_addr;
+
+       switch (dir) {
+       case FLOW_OFFLOAD_DIR_ORIGINAL:
+               addr = ip6h->daddr;
+               new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.src_v6;
+               ip6h->daddr = new_addr;
+               break;
+       case FLOW_OFFLOAD_DIR_REPLY:
+               addr = ip6h->saddr;
+               new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_v6;
+               ip6h->saddr = new_addr;
+               break;
+       default:
+               return -1;
+       }
+
+       return nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
+}
+
+static int nf_flow_nat_ipv6(const struct flow_offload *flow,
+                           struct sk_buff *skb,
+                           enum flow_offload_tuple_dir dir)
+{
+       struct ipv6hdr *ip6h = ipv6_hdr(skb);
+       unsigned int thoff = sizeof(*ip6h);
+
+       if (flow->flags & FLOW_OFFLOAD_SNAT &&
+           (nf_flow_snat_port(flow, skb, thoff, ip6h->nexthdr, dir) < 0 ||
+            nf_flow_snat_ipv6(flow, skb, ip6h, thoff, dir) < 0))
+               return -1;
+       if (flow->flags & FLOW_OFFLOAD_DNAT &&
+           (nf_flow_dnat_port(flow, skb, thoff, ip6h->nexthdr, dir) < 0 ||
+            nf_flow_dnat_ipv6(flow, skb, ip6h, thoff, dir) < 0))
+               return -1;
+
+       return 0;
+}
+
+static int nf_flow_tuple_ipv6(struct sk_buff *skb, const struct net_device *dev,
+                             struct flow_offload_tuple *tuple)
+{
+       struct flow_ports *ports;
+       struct ipv6hdr *ip6h;
+       unsigned int thoff;
+
+       if (!pskb_may_pull(skb, sizeof(*ip6h)))
+               return -1;
+
+       ip6h = ipv6_hdr(skb);
+
+       if (ip6h->nexthdr != IPPROTO_TCP &&
+           ip6h->nexthdr != IPPROTO_UDP)
+               return -1;
+
+       thoff = sizeof(*ip6h);
+       if (!pskb_may_pull(skb, thoff + sizeof(*ports)))
+               return -1;
+
+       ports = (struct flow_ports *)(skb_network_header(skb) + thoff);
+
+       tuple->src_v6           = ip6h->saddr;
+       tuple->dst_v6           = ip6h->daddr;
+       tuple->src_port         = ports->source;
+       tuple->dst_port         = ports->dest;
+       tuple->l3proto          = AF_INET6;
+       tuple->l4proto          = ip6h->nexthdr;
+       tuple->iifidx           = dev->ifindex;
+
+       return 0;
+}
+
+unsigned int
+nf_flow_offload_ipv6_hook(void *priv, struct sk_buff *skb,
+                         const struct nf_hook_state *state)
+{
+       struct flow_offload_tuple_rhash *tuplehash;
+       struct nf_flowtable *flow_table = priv;
+       struct flow_offload_tuple tuple = {};
+       enum flow_offload_tuple_dir dir;
+       struct flow_offload *flow;
+       struct net_device *outdev;
+       struct in6_addr *nexthop;
+       struct ipv6hdr *ip6h;
+       struct rt6_info *rt;
+
+       if (skb->protocol != htons(ETH_P_IPV6))
+               return NF_ACCEPT;
+
+       if (nf_flow_tuple_ipv6(skb, state->in, &tuple) < 0)
+               return NF_ACCEPT;
+
+       tuplehash = flow_offload_lookup(flow_table, &tuple);
+       if (tuplehash == NULL)
+               return NF_ACCEPT;
+
+       outdev = dev_get_by_index_rcu(state->net, tuplehash->tuple.oifidx);
+       if (!outdev)
+               return NF_ACCEPT;
+
+       dir = tuplehash->tuple.dir;
+       flow = container_of(tuplehash, struct flow_offload, tuplehash[dir]);
+       rt = (struct rt6_info *)flow->tuplehash[dir].tuple.dst_cache;
+
+       if (unlikely(nf_flow_exceeds_mtu(skb, flow->tuplehash[dir].tuple.mtu)))
+               return NF_ACCEPT;
+
+       if (skb_try_make_writable(skb, sizeof(*ip6h)))
+               return NF_DROP;
+
+       if (flow->flags & (FLOW_OFFLOAD_SNAT | FLOW_OFFLOAD_DNAT) &&
+           nf_flow_nat_ipv6(flow, skb, dir) < 0)
+               return NF_DROP;
+
+       flow->timeout = (u32)jiffies + NF_FLOW_TIMEOUT;
+       ip6h = ipv6_hdr(skb);
+       ip6h->hop_limit--;
+
+       skb->dev = outdev;
+       nexthop = rt6_nexthop(rt, &flow->tuplehash[!dir].tuple.src_v6);
+       neigh_xmit(NEIGH_ND_TABLE, outdev, nexthop, skb);
+
+       return NF_STOLEN;
+}
+EXPORT_SYMBOL_GPL(nf_flow_offload_ipv6_hook);