Add the rt linux 4.1.3-rt3 as base
[kvmfornfv.git] / kernel / net / ipv6 / ip6_input.c
diff --git a/kernel/net/ipv6/ip6_input.c b/kernel/net/ipv6/ip6_input.c
new file mode 100644 (file)
index 0000000..f2e464e
--- /dev/null
@@ -0,0 +1,363 @@
+/*
+ *     IPv6 input
+ *     Linux INET6 implementation
+ *
+ *     Authors:
+ *     Pedro Roque             <roque@di.fc.ul.pt>
+ *     Ian P. Morris           <I.P.Morris@soton.ac.uk>
+ *
+ *     Based in linux/net/ipv4/ip_input.c
+ *
+ *     This program is free software; you can redistribute it and/or
+ *      modify it under the terms of the GNU General Public License
+ *      as published by the Free Software Foundation; either version
+ *      2 of the License, or (at your option) any later version.
+ */
+/* Changes
+ *
+ *     Mitsuru KANDA @USAGI and
+ *     YOSHIFUJI Hideaki @USAGI: Remove ipv6_parse_exthdrs().
+ */
+
+#include <linux/errno.h>
+#include <linux/types.h>
+#include <linux/socket.h>
+#include <linux/sockios.h>
+#include <linux/net.h>
+#include <linux/netdevice.h>
+#include <linux/in6.h>
+#include <linux/icmpv6.h>
+#include <linux/mroute6.h>
+#include <linux/slab.h>
+
+#include <linux/netfilter.h>
+#include <linux/netfilter_ipv6.h>
+
+#include <net/sock.h>
+#include <net/snmp.h>
+
+#include <net/ipv6.h>
+#include <net/protocol.h>
+#include <net/transp_v6.h>
+#include <net/rawv6.h>
+#include <net/ndisc.h>
+#include <net/ip6_route.h>
+#include <net/addrconf.h>
+#include <net/xfrm.h>
+#include <net/inet_ecn.h>
+
+int ip6_rcv_finish(struct sock *sk, struct sk_buff *skb)
+{
+       if (sysctl_ip_early_demux && !skb_dst(skb) && skb->sk == NULL) {
+               const struct inet6_protocol *ipprot;
+
+               ipprot = rcu_dereference(inet6_protos[ipv6_hdr(skb)->nexthdr]);
+               if (ipprot && ipprot->early_demux)
+                       ipprot->early_demux(skb);
+       }
+       if (!skb_dst(skb))
+               ip6_route_input(skb);
+
+       return dst_input(skb);
+}
+
+int ipv6_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, struct net_device *orig_dev)
+{
+       const struct ipv6hdr *hdr;
+       u32 pkt_len;
+       struct inet6_dev *idev;
+       struct net *net = dev_net(skb->dev);
+
+       if (skb->pkt_type == PACKET_OTHERHOST) {
+               kfree_skb(skb);
+               return NET_RX_DROP;
+       }
+
+       rcu_read_lock();
+
+       idev = __in6_dev_get(skb->dev);
+
+       IP6_UPD_PO_STATS_BH(net, idev, IPSTATS_MIB_IN, skb->len);
+
+       if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL ||
+           !idev || unlikely(idev->cnf.disable_ipv6)) {
+               IP6_INC_STATS_BH(net, idev, IPSTATS_MIB_INDISCARDS);
+               goto drop;
+       }
+
+       memset(IP6CB(skb), 0, sizeof(struct inet6_skb_parm));
+
+       /*
+        * Store incoming device index. When the packet will
+        * be queued, we cannot refer to skb->dev anymore.
+        *
+        * BTW, when we send a packet for our own local address on a
+        * non-loopback interface (e.g. ethX), it is being delivered
+        * via the loopback interface (lo) here; skb->dev = loopback_dev.
+        * It, however, should be considered as if it is being
+        * arrived via the sending interface (ethX), because of the
+        * nature of scoping architecture. --yoshfuji
+        */
+       IP6CB(skb)->iif = skb_dst(skb) ? ip6_dst_idev(skb_dst(skb))->dev->ifindex : dev->ifindex;
+
+       if (unlikely(!pskb_may_pull(skb, sizeof(*hdr))))
+               goto err;
+
+       hdr = ipv6_hdr(skb);
+
+       if (hdr->version != 6)
+               goto err;
+
+       IP6_ADD_STATS_BH(dev_net(dev), idev,
+                        IPSTATS_MIB_NOECTPKTS +
+                               (ipv6_get_dsfield(hdr) & INET_ECN_MASK),
+                        max_t(unsigned short, 1, skb_shinfo(skb)->gso_segs));
+       /*
+        * RFC4291 2.5.3
+        * A packet received on an interface with a destination address
+        * of loopback must be dropped.
+        */
+       if (!(dev->flags & IFF_LOOPBACK) &&
+           ipv6_addr_loopback(&hdr->daddr))
+               goto err;
+
+       /* RFC4291 Errata ID: 3480
+        * Interface-Local scope spans only a single interface on a
+        * node and is useful only for loopback transmission of
+        * multicast.  Packets with interface-local scope received
+        * from another node must be discarded.
+        */
+       if (!(skb->pkt_type == PACKET_LOOPBACK ||
+             dev->flags & IFF_LOOPBACK) &&
+           ipv6_addr_is_multicast(&hdr->daddr) &&
+           IPV6_ADDR_MC_SCOPE(&hdr->daddr) == 1)
+               goto err;
+
+       /* RFC4291 2.7
+        * Nodes must not originate a packet to a multicast address whose scope
+        * field contains the reserved value 0; if such a packet is received, it
+        * must be silently dropped.
+        */
+       if (ipv6_addr_is_multicast(&hdr->daddr) &&
+           IPV6_ADDR_MC_SCOPE(&hdr->daddr) == 0)
+               goto err;
+
+       /*
+        * RFC4291 2.7
+        * Multicast addresses must not be used as source addresses in IPv6
+        * packets or appear in any Routing header.
+        */
+       if (ipv6_addr_is_multicast(&hdr->saddr))
+               goto err;
+
+       skb->transport_header = skb->network_header + sizeof(*hdr);
+       IP6CB(skb)->nhoff = offsetof(struct ipv6hdr, nexthdr);
+
+       pkt_len = ntohs(hdr->payload_len);
+
+       /* pkt_len may be zero if Jumbo payload option is present */
+       if (pkt_len || hdr->nexthdr != NEXTHDR_HOP) {
+               if (pkt_len + sizeof(struct ipv6hdr) > skb->len) {
+                       IP6_INC_STATS_BH(net,
+                                        idev, IPSTATS_MIB_INTRUNCATEDPKTS);
+                       goto drop;
+               }
+               if (pskb_trim_rcsum(skb, pkt_len + sizeof(struct ipv6hdr))) {
+                       IP6_INC_STATS_BH(net, idev, IPSTATS_MIB_INHDRERRORS);
+                       goto drop;
+               }
+               hdr = ipv6_hdr(skb);
+       }
+
+       if (hdr->nexthdr == NEXTHDR_HOP) {
+               if (ipv6_parse_hopopts(skb) < 0) {
+                       IP6_INC_STATS_BH(net, idev, IPSTATS_MIB_INHDRERRORS);
+                       rcu_read_unlock();
+                       return NET_RX_DROP;
+               }
+       }
+
+       rcu_read_unlock();
+
+       /* Must drop socket now because of tproxy. */
+       skb_orphan(skb);
+
+       return NF_HOOK(NFPROTO_IPV6, NF_INET_PRE_ROUTING, NULL, skb,
+                      dev, NULL,
+                      ip6_rcv_finish);
+err:
+       IP6_INC_STATS_BH(net, idev, IPSTATS_MIB_INHDRERRORS);
+drop:
+       rcu_read_unlock();
+       kfree_skb(skb);
+       return NET_RX_DROP;
+}
+
+/*
+ *     Deliver the packet to the host
+ */
+
+
+static int ip6_input_finish(struct sock *sk, struct sk_buff *skb)
+{
+       struct net *net = dev_net(skb_dst(skb)->dev);
+       const struct inet6_protocol *ipprot;
+       struct inet6_dev *idev;
+       unsigned int nhoff;
+       int nexthdr;
+       bool raw;
+
+       /*
+        *      Parse extension headers
+        */
+
+       rcu_read_lock();
+resubmit:
+       idev = ip6_dst_idev(skb_dst(skb));
+       if (!pskb_pull(skb, skb_transport_offset(skb)))
+               goto discard;
+       nhoff = IP6CB(skb)->nhoff;
+       nexthdr = skb_network_header(skb)[nhoff];
+
+       raw = raw6_local_deliver(skb, nexthdr);
+       ipprot = rcu_dereference(inet6_protos[nexthdr]);
+       if (ipprot) {
+               int ret;
+
+               if (ipprot->flags & INET6_PROTO_FINAL) {
+                       const struct ipv6hdr *hdr;
+
+                       /* Free reference early: we don't need it any more,
+                          and it may hold ip_conntrack module loaded
+                          indefinitely. */
+                       nf_reset(skb);
+
+                       skb_postpull_rcsum(skb, skb_network_header(skb),
+                                          skb_network_header_len(skb));
+                       hdr = ipv6_hdr(skb);
+                       if (ipv6_addr_is_multicast(&hdr->daddr) &&
+                           !ipv6_chk_mcast_addr(skb->dev, &hdr->daddr,
+                           &hdr->saddr) &&
+                           !ipv6_is_mld(skb, nexthdr, skb_network_header_len(skb)))
+                               goto discard;
+               }
+               if (!(ipprot->flags & INET6_PROTO_NOPOLICY) &&
+                   !xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
+                       goto discard;
+
+               ret = ipprot->handler(skb);
+               if (ret > 0)
+                       goto resubmit;
+               else if (ret == 0)
+                       IP6_INC_STATS_BH(net, idev, IPSTATS_MIB_INDELIVERS);
+       } else {
+               if (!raw) {
+                       if (xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb)) {
+                               IP6_INC_STATS_BH(net, idev,
+                                                IPSTATS_MIB_INUNKNOWNPROTOS);
+                               icmpv6_send(skb, ICMPV6_PARAMPROB,
+                                           ICMPV6_UNK_NEXTHDR, nhoff);
+                       }
+                       kfree_skb(skb);
+               } else {
+                       IP6_INC_STATS_BH(net, idev, IPSTATS_MIB_INDELIVERS);
+                       consume_skb(skb);
+               }
+       }
+       rcu_read_unlock();
+       return 0;
+
+discard:
+       IP6_INC_STATS_BH(net, idev, IPSTATS_MIB_INDISCARDS);
+       rcu_read_unlock();
+       kfree_skb(skb);
+       return 0;
+}
+
+
+int ip6_input(struct sk_buff *skb)
+{
+       return NF_HOOK(NFPROTO_IPV6, NF_INET_LOCAL_IN, NULL, skb,
+                      skb->dev, NULL,
+                      ip6_input_finish);
+}
+
+int ip6_mc_input(struct sk_buff *skb)
+{
+       const struct ipv6hdr *hdr;
+       bool deliver;
+
+       IP6_UPD_PO_STATS_BH(dev_net(skb_dst(skb)->dev),
+                        ip6_dst_idev(skb_dst(skb)), IPSTATS_MIB_INMCAST,
+                        skb->len);
+
+       hdr = ipv6_hdr(skb);
+       deliver = ipv6_chk_mcast_addr(skb->dev, &hdr->daddr, NULL);
+
+#ifdef CONFIG_IPV6_MROUTE
+       /*
+        *      IPv6 multicast router mode is now supported ;)
+        */
+       if (dev_net(skb->dev)->ipv6.devconf_all->mc_forwarding &&
+           !(ipv6_addr_type(&hdr->daddr) &
+             (IPV6_ADDR_LOOPBACK|IPV6_ADDR_LINKLOCAL)) &&
+           likely(!(IP6CB(skb)->flags & IP6SKB_FORWARDED))) {
+               /*
+                * Okay, we try to forward - split and duplicate
+                * packets.
+                */
+               struct sk_buff *skb2;
+               struct inet6_skb_parm *opt = IP6CB(skb);
+
+               /* Check for MLD */
+               if (unlikely(opt->flags & IP6SKB_ROUTERALERT)) {
+                       /* Check if this is a mld message */
+                       u8 nexthdr = hdr->nexthdr;
+                       __be16 frag_off;
+                       int offset;
+
+                       /* Check if the value of Router Alert
+                        * is for MLD (0x0000).
+                        */
+                       if (opt->ra == htons(IPV6_OPT_ROUTERALERT_MLD)) {
+                               deliver = false;
+
+                               if (!ipv6_ext_hdr(nexthdr)) {
+                                       /* BUG */
+                                       goto out;
+                               }
+                               offset = ipv6_skip_exthdr(skb, sizeof(*hdr),
+                                                         &nexthdr, &frag_off);
+                               if (offset < 0)
+                                       goto out;
+
+                               if (!ipv6_is_mld(skb, nexthdr, offset))
+                                       goto out;
+
+                               deliver = true;
+                       }
+                       /* unknown RA - process it normally */
+               }
+
+               if (deliver)
+                       skb2 = skb_clone(skb, GFP_ATOMIC);
+               else {
+                       skb2 = skb;
+                       skb = NULL;
+               }
+
+               if (skb2) {
+                       ip6_mr_input(skb2);
+               }
+       }
+out:
+#endif
+       if (likely(deliver))
+               ip6_input(skb);
+       else {
+               /* discard */
+               kfree_skb(skb);
+       }
+
+       return 0;
+}