Add the rt linux 4.1.3-rt3 as base
[kvmfornfv.git] / kernel / net / openvswitch / flow.c
diff --git a/kernel/net/openvswitch/flow.c b/kernel/net/openvswitch/flow.c
new file mode 100644 (file)
index 0000000..2dacc7b
--- /dev/null
@@ -0,0 +1,730 @@
+/*
+ * Copyright (c) 2007-2014 Nicira, Inc.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of version 2 of the GNU General Public
+ * License as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
+ * 02110-1301, USA
+ */
+
+#include <linux/uaccess.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/if_ether.h>
+#include <linux/if_vlan.h>
+#include <net/llc_pdu.h>
+#include <linux/kernel.h>
+#include <linux/jhash.h>
+#include <linux/jiffies.h>
+#include <linux/llc.h>
+#include <linux/module.h>
+#include <linux/in.h>
+#include <linux/rcupdate.h>
+#include <linux/if_arp.h>
+#include <linux/ip.h>
+#include <linux/ipv6.h>
+#include <linux/mpls.h>
+#include <linux/sctp.h>
+#include <linux/smp.h>
+#include <linux/tcp.h>
+#include <linux/udp.h>
+#include <linux/icmp.h>
+#include <linux/icmpv6.h>
+#include <linux/rculist.h>
+#include <net/ip.h>
+#include <net/ip_tunnels.h>
+#include <net/ipv6.h>
+#include <net/mpls.h>
+#include <net/ndisc.h>
+
+#include "datapath.h"
+#include "flow.h"
+#include "flow_netlink.h"
+
+u64 ovs_flow_used_time(unsigned long flow_jiffies)
+{
+       struct timespec cur_ts;
+       u64 cur_ms, idle_ms;
+
+       ktime_get_ts(&cur_ts);
+       idle_ms = jiffies_to_msecs(jiffies - flow_jiffies);
+       cur_ms = (u64)cur_ts.tv_sec * MSEC_PER_SEC +
+                cur_ts.tv_nsec / NSEC_PER_MSEC;
+
+       return cur_ms - idle_ms;
+}
+
+#define TCP_FLAGS_BE16(tp) (*(__be16 *)&tcp_flag_word(tp) & htons(0x0FFF))
+
+void ovs_flow_stats_update(struct sw_flow *flow, __be16 tcp_flags,
+                          const struct sk_buff *skb)
+{
+       struct flow_stats *stats;
+       int node = numa_node_id();
+       int len = skb->len + (skb_vlan_tag_present(skb) ? VLAN_HLEN : 0);
+
+       stats = rcu_dereference(flow->stats[node]);
+
+       /* Check if already have node-specific stats. */
+       if (likely(stats)) {
+               spin_lock(&stats->lock);
+               /* Mark if we write on the pre-allocated stats. */
+               if (node == 0 && unlikely(flow->stats_last_writer != node))
+                       flow->stats_last_writer = node;
+       } else {
+               stats = rcu_dereference(flow->stats[0]); /* Pre-allocated. */
+               spin_lock(&stats->lock);
+
+               /* If the current NUMA-node is the only writer on the
+                * pre-allocated stats keep using them.
+                */
+               if (unlikely(flow->stats_last_writer != node)) {
+                       /* A previous locker may have already allocated the
+                        * stats, so we need to check again.  If node-specific
+                        * stats were already allocated, we update the pre-
+                        * allocated stats as we have already locked them.
+                        */
+                       if (likely(flow->stats_last_writer != NUMA_NO_NODE)
+                           && likely(!rcu_access_pointer(flow->stats[node]))) {
+                               /* Try to allocate node-specific stats. */
+                               struct flow_stats *new_stats;
+
+                               new_stats =
+                                       kmem_cache_alloc_node(flow_stats_cache,
+                                                             GFP_NOWAIT |
+                                                             __GFP_THISNODE |
+                                                             __GFP_NOWARN |
+                                                             __GFP_NOMEMALLOC,
+                                                             node);
+                               if (likely(new_stats)) {
+                                       new_stats->used = jiffies;
+                                       new_stats->packet_count = 1;
+                                       new_stats->byte_count = len;
+                                       new_stats->tcp_flags = tcp_flags;
+                                       spin_lock_init(&new_stats->lock);
+
+                                       rcu_assign_pointer(flow->stats[node],
+                                                          new_stats);
+                                       goto unlock;
+                               }
+                       }
+                       flow->stats_last_writer = node;
+               }
+       }
+
+       stats->used = jiffies;
+       stats->packet_count++;
+       stats->byte_count += len;
+       stats->tcp_flags |= tcp_flags;
+unlock:
+       spin_unlock(&stats->lock);
+}
+
+/* Must be called with rcu_read_lock or ovs_mutex. */
+void ovs_flow_stats_get(const struct sw_flow *flow,
+                       struct ovs_flow_stats *ovs_stats,
+                       unsigned long *used, __be16 *tcp_flags)
+{
+       int node;
+
+       *used = 0;
+       *tcp_flags = 0;
+       memset(ovs_stats, 0, sizeof(*ovs_stats));
+
+       for_each_node(node) {
+               struct flow_stats *stats = rcu_dereference_ovsl(flow->stats[node]);
+
+               if (stats) {
+                       /* Local CPU may write on non-local stats, so we must
+                        * block bottom-halves here.
+                        */
+                       spin_lock_bh(&stats->lock);
+                       if (!*used || time_after(stats->used, *used))
+                               *used = stats->used;
+                       *tcp_flags |= stats->tcp_flags;
+                       ovs_stats->n_packets += stats->packet_count;
+                       ovs_stats->n_bytes += stats->byte_count;
+                       spin_unlock_bh(&stats->lock);
+               }
+       }
+}
+
+/* Called with ovs_mutex. */
+void ovs_flow_stats_clear(struct sw_flow *flow)
+{
+       int node;
+
+       for_each_node(node) {
+               struct flow_stats *stats = ovsl_dereference(flow->stats[node]);
+
+               if (stats) {
+                       spin_lock_bh(&stats->lock);
+                       stats->used = 0;
+                       stats->packet_count = 0;
+                       stats->byte_count = 0;
+                       stats->tcp_flags = 0;
+                       spin_unlock_bh(&stats->lock);
+               }
+       }
+}
+
+static int check_header(struct sk_buff *skb, int len)
+{
+       if (unlikely(skb->len < len))
+               return -EINVAL;
+       if (unlikely(!pskb_may_pull(skb, len)))
+               return -ENOMEM;
+       return 0;
+}
+
+static bool arphdr_ok(struct sk_buff *skb)
+{
+       return pskb_may_pull(skb, skb_network_offset(skb) +
+                                 sizeof(struct arp_eth_header));
+}
+
+static int check_iphdr(struct sk_buff *skb)
+{
+       unsigned int nh_ofs = skb_network_offset(skb);
+       unsigned int ip_len;
+       int err;
+
+       err = check_header(skb, nh_ofs + sizeof(struct iphdr));
+       if (unlikely(err))
+               return err;
+
+       ip_len = ip_hdrlen(skb);
+       if (unlikely(ip_len < sizeof(struct iphdr) ||
+                    skb->len < nh_ofs + ip_len))
+               return -EINVAL;
+
+       skb_set_transport_header(skb, nh_ofs + ip_len);
+       return 0;
+}
+
+static bool tcphdr_ok(struct sk_buff *skb)
+{
+       int th_ofs = skb_transport_offset(skb);
+       int tcp_len;
+
+       if (unlikely(!pskb_may_pull(skb, th_ofs + sizeof(struct tcphdr))))
+               return false;
+
+       tcp_len = tcp_hdrlen(skb);
+       if (unlikely(tcp_len < sizeof(struct tcphdr) ||
+                    skb->len < th_ofs + tcp_len))
+               return false;
+
+       return true;
+}
+
+static bool udphdr_ok(struct sk_buff *skb)
+{
+       return pskb_may_pull(skb, skb_transport_offset(skb) +
+                                 sizeof(struct udphdr));
+}
+
+static bool sctphdr_ok(struct sk_buff *skb)
+{
+       return pskb_may_pull(skb, skb_transport_offset(skb) +
+                                 sizeof(struct sctphdr));
+}
+
+static bool icmphdr_ok(struct sk_buff *skb)
+{
+       return pskb_may_pull(skb, skb_transport_offset(skb) +
+                                 sizeof(struct icmphdr));
+}
+
+static int parse_ipv6hdr(struct sk_buff *skb, struct sw_flow_key *key)
+{
+       unsigned int nh_ofs = skb_network_offset(skb);
+       unsigned int nh_len;
+       int payload_ofs;
+       struct ipv6hdr *nh;
+       uint8_t nexthdr;
+       __be16 frag_off;
+       int err;
+
+       err = check_header(skb, nh_ofs + sizeof(*nh));
+       if (unlikely(err))
+               return err;
+
+       nh = ipv6_hdr(skb);
+       nexthdr = nh->nexthdr;
+       payload_ofs = (u8 *)(nh + 1) - skb->data;
+
+       key->ip.proto = NEXTHDR_NONE;
+       key->ip.tos = ipv6_get_dsfield(nh);
+       key->ip.ttl = nh->hop_limit;
+       key->ipv6.label = *(__be32 *)nh & htonl(IPV6_FLOWINFO_FLOWLABEL);
+       key->ipv6.addr.src = nh->saddr;
+       key->ipv6.addr.dst = nh->daddr;
+
+       payload_ofs = ipv6_skip_exthdr(skb, payload_ofs, &nexthdr, &frag_off);
+       if (unlikely(payload_ofs < 0))
+               return -EINVAL;
+
+       if (frag_off) {
+               if (frag_off & htons(~0x7))
+                       key->ip.frag = OVS_FRAG_TYPE_LATER;
+               else
+                       key->ip.frag = OVS_FRAG_TYPE_FIRST;
+       } else {
+               key->ip.frag = OVS_FRAG_TYPE_NONE;
+       }
+
+       nh_len = payload_ofs - nh_ofs;
+       skb_set_transport_header(skb, nh_ofs + nh_len);
+       key->ip.proto = nexthdr;
+       return nh_len;
+}
+
+static bool icmp6hdr_ok(struct sk_buff *skb)
+{
+       return pskb_may_pull(skb, skb_transport_offset(skb) +
+                                 sizeof(struct icmp6hdr));
+}
+
+static int parse_vlan(struct sk_buff *skb, struct sw_flow_key *key)
+{
+       struct qtag_prefix {
+               __be16 eth_type; /* ETH_P_8021Q */
+               __be16 tci;
+       };
+       struct qtag_prefix *qp;
+
+       if (unlikely(skb->len < sizeof(struct qtag_prefix) + sizeof(__be16)))
+               return 0;
+
+       if (unlikely(!pskb_may_pull(skb, sizeof(struct qtag_prefix) +
+                                        sizeof(__be16))))
+               return -ENOMEM;
+
+       qp = (struct qtag_prefix *) skb->data;
+       key->eth.tci = qp->tci | htons(VLAN_TAG_PRESENT);
+       __skb_pull(skb, sizeof(struct qtag_prefix));
+
+       return 0;
+}
+
+static __be16 parse_ethertype(struct sk_buff *skb)
+{
+       struct llc_snap_hdr {
+               u8  dsap;  /* Always 0xAA */
+               u8  ssap;  /* Always 0xAA */
+               u8  ctrl;
+               u8  oui[3];
+               __be16 ethertype;
+       };
+       struct llc_snap_hdr *llc;
+       __be16 proto;
+
+       proto = *(__be16 *) skb->data;
+       __skb_pull(skb, sizeof(__be16));
+
+       if (ntohs(proto) >= ETH_P_802_3_MIN)
+               return proto;
+
+       if (skb->len < sizeof(struct llc_snap_hdr))
+               return htons(ETH_P_802_2);
+
+       if (unlikely(!pskb_may_pull(skb, sizeof(struct llc_snap_hdr))))
+               return htons(0);
+
+       llc = (struct llc_snap_hdr *) skb->data;
+       if (llc->dsap != LLC_SAP_SNAP ||
+           llc->ssap != LLC_SAP_SNAP ||
+           (llc->oui[0] | llc->oui[1] | llc->oui[2]) != 0)
+               return htons(ETH_P_802_2);
+
+       __skb_pull(skb, sizeof(struct llc_snap_hdr));
+
+       if (ntohs(llc->ethertype) >= ETH_P_802_3_MIN)
+               return llc->ethertype;
+
+       return htons(ETH_P_802_2);
+}
+
+static int parse_icmpv6(struct sk_buff *skb, struct sw_flow_key *key,
+                       int nh_len)
+{
+       struct icmp6hdr *icmp = icmp6_hdr(skb);
+
+       /* The ICMPv6 type and code fields use the 16-bit transport port
+        * fields, so we need to store them in 16-bit network byte order.
+        */
+       key->tp.src = htons(icmp->icmp6_type);
+       key->tp.dst = htons(icmp->icmp6_code);
+       memset(&key->ipv6.nd, 0, sizeof(key->ipv6.nd));
+
+       if (icmp->icmp6_code == 0 &&
+           (icmp->icmp6_type == NDISC_NEIGHBOUR_SOLICITATION ||
+            icmp->icmp6_type == NDISC_NEIGHBOUR_ADVERTISEMENT)) {
+               int icmp_len = skb->len - skb_transport_offset(skb);
+               struct nd_msg *nd;
+               int offset;
+
+               /* In order to process neighbor discovery options, we need the
+                * entire packet.
+                */
+               if (unlikely(icmp_len < sizeof(*nd)))
+                       return 0;
+
+               if (unlikely(skb_linearize(skb)))
+                       return -ENOMEM;
+
+               nd = (struct nd_msg *)skb_transport_header(skb);
+               key->ipv6.nd.target = nd->target;
+
+               icmp_len -= sizeof(*nd);
+               offset = 0;
+               while (icmp_len >= 8) {
+                       struct nd_opt_hdr *nd_opt =
+                                (struct nd_opt_hdr *)(nd->opt + offset);
+                       int opt_len = nd_opt->nd_opt_len * 8;
+
+                       if (unlikely(!opt_len || opt_len > icmp_len))
+                               return 0;
+
+                       /* Store the link layer address if the appropriate
+                        * option is provided.  It is considered an error if
+                        * the same link layer option is specified twice.
+                        */
+                       if (nd_opt->nd_opt_type == ND_OPT_SOURCE_LL_ADDR
+                           && opt_len == 8) {
+                               if (unlikely(!is_zero_ether_addr(key->ipv6.nd.sll)))
+                                       goto invalid;
+                               ether_addr_copy(key->ipv6.nd.sll,
+                                               &nd->opt[offset+sizeof(*nd_opt)]);
+                       } else if (nd_opt->nd_opt_type == ND_OPT_TARGET_LL_ADDR
+                                  && opt_len == 8) {
+                               if (unlikely(!is_zero_ether_addr(key->ipv6.nd.tll)))
+                                       goto invalid;
+                               ether_addr_copy(key->ipv6.nd.tll,
+                                               &nd->opt[offset+sizeof(*nd_opt)]);
+                       }
+
+                       icmp_len -= opt_len;
+                       offset += opt_len;
+               }
+       }
+
+       return 0;
+
+invalid:
+       memset(&key->ipv6.nd.target, 0, sizeof(key->ipv6.nd.target));
+       memset(key->ipv6.nd.sll, 0, sizeof(key->ipv6.nd.sll));
+       memset(key->ipv6.nd.tll, 0, sizeof(key->ipv6.nd.tll));
+
+       return 0;
+}
+
+/**
+ * key_extract - extracts a flow key from an Ethernet frame.
+ * @skb: sk_buff that contains the frame, with skb->data pointing to the
+ * Ethernet header
+ * @key: output flow key
+ *
+ * The caller must ensure that skb->len >= ETH_HLEN.
+ *
+ * Returns 0 if successful, otherwise a negative errno value.
+ *
+ * Initializes @skb header pointers as follows:
+ *
+ *    - skb->mac_header: the Ethernet header.
+ *
+ *    - skb->network_header: just past the Ethernet header, or just past the
+ *      VLAN header, to the first byte of the Ethernet payload.
+ *
+ *    - skb->transport_header: If key->eth.type is ETH_P_IP or ETH_P_IPV6
+ *      on output, then just past the IP header, if one is present and
+ *      of a correct length, otherwise the same as skb->network_header.
+ *      For other key->eth.type values it is left untouched.
+ */
+static int key_extract(struct sk_buff *skb, struct sw_flow_key *key)
+{
+       int error;
+       struct ethhdr *eth;
+
+       /* Flags are always used as part of stats */
+       key->tp.flags = 0;
+
+       skb_reset_mac_header(skb);
+
+       /* Link layer.  We are guaranteed to have at least the 14 byte Ethernet
+        * header in the linear data area.
+        */
+       eth = eth_hdr(skb);
+       ether_addr_copy(key->eth.src, eth->h_source);
+       ether_addr_copy(key->eth.dst, eth->h_dest);
+
+       __skb_pull(skb, 2 * ETH_ALEN);
+       /* We are going to push all headers that we pull, so no need to
+        * update skb->csum here.
+        */
+
+       key->eth.tci = 0;
+       if (skb_vlan_tag_present(skb))
+               key->eth.tci = htons(skb->vlan_tci);
+       else if (eth->h_proto == htons(ETH_P_8021Q))
+               if (unlikely(parse_vlan(skb, key)))
+                       return -ENOMEM;
+
+       key->eth.type = parse_ethertype(skb);
+       if (unlikely(key->eth.type == htons(0)))
+               return -ENOMEM;
+
+       skb_reset_network_header(skb);
+       skb_reset_mac_len(skb);
+       __skb_push(skb, skb->data - skb_mac_header(skb));
+
+       /* Network layer. */
+       if (key->eth.type == htons(ETH_P_IP)) {
+               struct iphdr *nh;
+               __be16 offset;
+
+               error = check_iphdr(skb);
+               if (unlikely(error)) {
+                       memset(&key->ip, 0, sizeof(key->ip));
+                       memset(&key->ipv4, 0, sizeof(key->ipv4));
+                       if (error == -EINVAL) {
+                               skb->transport_header = skb->network_header;
+                               error = 0;
+                       }
+                       return error;
+               }
+
+               nh = ip_hdr(skb);
+               key->ipv4.addr.src = nh->saddr;
+               key->ipv4.addr.dst = nh->daddr;
+
+               key->ip.proto = nh->protocol;
+               key->ip.tos = nh->tos;
+               key->ip.ttl = nh->ttl;
+
+               offset = nh->frag_off & htons(IP_OFFSET);
+               if (offset) {
+                       key->ip.frag = OVS_FRAG_TYPE_LATER;
+                       return 0;
+               }
+               if (nh->frag_off & htons(IP_MF) ||
+                       skb_shinfo(skb)->gso_type & SKB_GSO_UDP)
+                       key->ip.frag = OVS_FRAG_TYPE_FIRST;
+               else
+                       key->ip.frag = OVS_FRAG_TYPE_NONE;
+
+               /* Transport layer. */
+               if (key->ip.proto == IPPROTO_TCP) {
+                       if (tcphdr_ok(skb)) {
+                               struct tcphdr *tcp = tcp_hdr(skb);
+                               key->tp.src = tcp->source;
+                               key->tp.dst = tcp->dest;
+                               key->tp.flags = TCP_FLAGS_BE16(tcp);
+                       } else {
+                               memset(&key->tp, 0, sizeof(key->tp));
+                       }
+
+               } else if (key->ip.proto == IPPROTO_UDP) {
+                       if (udphdr_ok(skb)) {
+                               struct udphdr *udp = udp_hdr(skb);
+                               key->tp.src = udp->source;
+                               key->tp.dst = udp->dest;
+                       } else {
+                               memset(&key->tp, 0, sizeof(key->tp));
+                       }
+               } else if (key->ip.proto == IPPROTO_SCTP) {
+                       if (sctphdr_ok(skb)) {
+                               struct sctphdr *sctp = sctp_hdr(skb);
+                               key->tp.src = sctp->source;
+                               key->tp.dst = sctp->dest;
+                       } else {
+                               memset(&key->tp, 0, sizeof(key->tp));
+                       }
+               } else if (key->ip.proto == IPPROTO_ICMP) {
+                       if (icmphdr_ok(skb)) {
+                               struct icmphdr *icmp = icmp_hdr(skb);
+                               /* The ICMP type and code fields use the 16-bit
+                                * transport port fields, so we need to store
+                                * them in 16-bit network byte order. */
+                               key->tp.src = htons(icmp->type);
+                               key->tp.dst = htons(icmp->code);
+                       } else {
+                               memset(&key->tp, 0, sizeof(key->tp));
+                       }
+               }
+
+       } else if (key->eth.type == htons(ETH_P_ARP) ||
+                  key->eth.type == htons(ETH_P_RARP)) {
+               struct arp_eth_header *arp;
+               bool arp_available = arphdr_ok(skb);
+
+               arp = (struct arp_eth_header *)skb_network_header(skb);
+
+               if (arp_available &&
+                   arp->ar_hrd == htons(ARPHRD_ETHER) &&
+                   arp->ar_pro == htons(ETH_P_IP) &&
+                   arp->ar_hln == ETH_ALEN &&
+                   arp->ar_pln == 4) {
+
+                       /* We only match on the lower 8 bits of the opcode. */
+                       if (ntohs(arp->ar_op) <= 0xff)
+                               key->ip.proto = ntohs(arp->ar_op);
+                       else
+                               key->ip.proto = 0;
+
+                       memcpy(&key->ipv4.addr.src, arp->ar_sip, sizeof(key->ipv4.addr.src));
+                       memcpy(&key->ipv4.addr.dst, arp->ar_tip, sizeof(key->ipv4.addr.dst));
+                       ether_addr_copy(key->ipv4.arp.sha, arp->ar_sha);
+                       ether_addr_copy(key->ipv4.arp.tha, arp->ar_tha);
+               } else {
+                       memset(&key->ip, 0, sizeof(key->ip));
+                       memset(&key->ipv4, 0, sizeof(key->ipv4));
+               }
+       } else if (eth_p_mpls(key->eth.type)) {
+               size_t stack_len = MPLS_HLEN;
+
+               /* In the presence of an MPLS label stack the end of the L2
+                * header and the beginning of the L3 header differ.
+                *
+                * Advance network_header to the beginning of the L3
+                * header. mac_len corresponds to the end of the L2 header.
+                */
+               while (1) {
+                       __be32 lse;
+
+                       error = check_header(skb, skb->mac_len + stack_len);
+                       if (unlikely(error))
+                               return 0;
+
+                       memcpy(&lse, skb_network_header(skb), MPLS_HLEN);
+
+                       if (stack_len == MPLS_HLEN)
+                               memcpy(&key->mpls.top_lse, &lse, MPLS_HLEN);
+
+                       skb_set_network_header(skb, skb->mac_len + stack_len);
+                       if (lse & htonl(MPLS_LS_S_MASK))
+                               break;
+
+                       stack_len += MPLS_HLEN;
+               }
+       } else if (key->eth.type == htons(ETH_P_IPV6)) {
+               int nh_len;             /* IPv6 Header + Extensions */
+
+               nh_len = parse_ipv6hdr(skb, key);
+               if (unlikely(nh_len < 0)) {
+                       memset(&key->ip, 0, sizeof(key->ip));
+                       memset(&key->ipv6.addr, 0, sizeof(key->ipv6.addr));
+                       if (nh_len == -EINVAL) {
+                               skb->transport_header = skb->network_header;
+                               error = 0;
+                       } else {
+                               error = nh_len;
+                       }
+                       return error;
+               }
+
+               if (key->ip.frag == OVS_FRAG_TYPE_LATER)
+                       return 0;
+               if (skb_shinfo(skb)->gso_type & SKB_GSO_UDP)
+                       key->ip.frag = OVS_FRAG_TYPE_FIRST;
+
+               /* Transport layer. */
+               if (key->ip.proto == NEXTHDR_TCP) {
+                       if (tcphdr_ok(skb)) {
+                               struct tcphdr *tcp = tcp_hdr(skb);
+                               key->tp.src = tcp->source;
+                               key->tp.dst = tcp->dest;
+                               key->tp.flags = TCP_FLAGS_BE16(tcp);
+                       } else {
+                               memset(&key->tp, 0, sizeof(key->tp));
+                       }
+               } else if (key->ip.proto == NEXTHDR_UDP) {
+                       if (udphdr_ok(skb)) {
+                               struct udphdr *udp = udp_hdr(skb);
+                               key->tp.src = udp->source;
+                               key->tp.dst = udp->dest;
+                       } else {
+                               memset(&key->tp, 0, sizeof(key->tp));
+                       }
+               } else if (key->ip.proto == NEXTHDR_SCTP) {
+                       if (sctphdr_ok(skb)) {
+                               struct sctphdr *sctp = sctp_hdr(skb);
+                               key->tp.src = sctp->source;
+                               key->tp.dst = sctp->dest;
+                       } else {
+                               memset(&key->tp, 0, sizeof(key->tp));
+                       }
+               } else if (key->ip.proto == NEXTHDR_ICMP) {
+                       if (icmp6hdr_ok(skb)) {
+                               error = parse_icmpv6(skb, key, nh_len);
+                               if (error)
+                                       return error;
+                       } else {
+                               memset(&key->tp, 0, sizeof(key->tp));
+                       }
+               }
+       }
+       return 0;
+}
+
+int ovs_flow_key_update(struct sk_buff *skb, struct sw_flow_key *key)
+{
+       return key_extract(skb, key);
+}
+
+int ovs_flow_key_extract(const struct ovs_tunnel_info *tun_info,
+                        struct sk_buff *skb, struct sw_flow_key *key)
+{
+       /* Extract metadata from packet. */
+       if (tun_info) {
+               memcpy(&key->tun_key, &tun_info->tunnel, sizeof(key->tun_key));
+
+               if (tun_info->options) {
+                       BUILD_BUG_ON((1 << (sizeof(tun_info->options_len) *
+                                                  8)) - 1
+                                       > sizeof(key->tun_opts));
+                       memcpy(TUN_METADATA_OPTS(key, tun_info->options_len),
+                              tun_info->options, tun_info->options_len);
+                       key->tun_opts_len = tun_info->options_len;
+               } else {
+                       key->tun_opts_len = 0;
+               }
+       } else  {
+               key->tun_opts_len = 0;
+               memset(&key->tun_key, 0, sizeof(key->tun_key));
+       }
+
+       key->phy.priority = skb->priority;
+       key->phy.in_port = OVS_CB(skb)->input_vport->port_no;
+       key->phy.skb_mark = skb->mark;
+       key->ovs_flow_hash = 0;
+       key->recirc_id = 0;
+
+       return key_extract(skb, key);
+}
+
+int ovs_flow_key_extract_userspace(const struct nlattr *attr,
+                                  struct sk_buff *skb,
+                                  struct sw_flow_key *key, bool log)
+{
+       int err;
+
+       memset(key, 0, OVS_SW_FLOW_KEY_METADATA_SIZE);
+
+       /* Extract metadata from netlink attributes. */
+       err = ovs_nla_get_flow_metadata(attr, key, log);
+       if (err)
+               return err;
+
+       return key_extract(skb, key);
+}