flow.c

		};

		if (type > OVS_TUNNEL_KEY_ATTR_MAX ||
			ovs_tunnel_key_lens[type] != nla_len(a))
			return -EINVAL;

		switch (type) {
		case OVS_TUNNEL_KEY_ATTR_ID:
			tun_key->tun_id = nla_get_be64(a);
			tun_key->tun_flags |= TUNNEL_KEY;
			break;
		case OVS_TUNNEL_KEY_ATTR_IPV4_SRC:
			tun_key->ipv4_src = nla_get_be32(a);
			break;
		case OVS_TUNNEL_KEY_ATTR_IPV4_DST:
			tun_key->ipv4_dst = nla_get_be32(a);
			break;
		case OVS_TUNNEL_KEY_ATTR_TOS:
			tun_key->ipv4_tos = nla_get_u8(a);
			break;
		case OVS_TUNNEL_KEY_ATTR_TTL:
			tun_key->ipv4_ttl = nla_get_u8(a);
			ttl = true;
			break;
		case OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT:
			tun_key->tun_flags |= TUNNEL_DONT_FRAGMENT;
			break;
		case OVS_TUNNEL_KEY_ATTR_CSUM:
			tun_key->tun_flags |= TUNNEL_CSUM;
			break;
		default:
			return -EINVAL;

		}
	}
	if (rem > 0)
		return -EINVAL;

	if (!tun_key->ipv4_dst)
		return -EINVAL;

	if (!ttl)
		return -EINVAL;

	return 0;
}

int ovs_ipv4_tun_to_nlattr(struct sk_buff *skb,
			   const struct ovs_key_ipv4_tunnel *tun_key)
{
	struct nlattr *nla;

	nla = nla_nest_start(skb, OVS_KEY_ATTR_TUNNEL);
	if (!nla)
		return -EMSGSIZE;

	if (tun_key->tun_flags & TUNNEL_KEY &&
	    nla_put_be64(skb, OVS_TUNNEL_KEY_ATTR_ID, tun_key->tun_id))
		return -EMSGSIZE;
	if (tun_key->ipv4_src &&
	    nla_put_be32(skb, OVS_TUNNEL_KEY_ATTR_IPV4_SRC, tun_key->ipv4_src))
		return -EMSGSIZE;
	if (nla_put_be32(skb, OVS_TUNNEL_KEY_ATTR_IPV4_DST, tun_key->ipv4_dst))
		return -EMSGSIZE;
	if (tun_key->ipv4_tos &&
	    nla_put_u8(skb, OVS_TUNNEL_KEY_ATTR_TOS, tun_key->ipv4_tos))
		return -EMSGSIZE;
	if (nla_put_u8(skb, OVS_TUNNEL_KEY_ATTR_TTL, tun_key->ipv4_ttl))
		return -EMSGSIZE;
	if ((tun_key->tun_flags & TUNNEL_DONT_FRAGMENT) &&
		nla_put_flag(skb, OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT))
		return -EMSGSIZE;
	if ((tun_key->tun_flags & TUNNEL_CSUM) &&
		nla_put_flag(skb, OVS_TUNNEL_KEY_ATTR_CSUM))
		return -EMSGSIZE;

	nla_nest_end(skb, nla);
	return 0;
}

/**
 * ovs_flow_from_nlattrs - parses Netlink attributes into a flow key.
 * @swkey: receives the extracted flow key.
 * @key_lenp: number of bytes used in @swkey.
 * @attr: Netlink attribute holding nested %OVS_KEY_ATTR_* Netlink attribute
 * sequence.
 */
int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp,
		      const struct nlattr *attr)
{
	const struct nlattr *a[OVS_KEY_ATTR_MAX + 1];
	const struct ovs_key_ethernet *eth_key;
	int key_len;
	u32 attrs;
	int err;

	memset(swkey, 0, sizeof(struct sw_flow_key));
	key_len = SW_FLOW_KEY_OFFSET(eth);

	err = parse_flow_nlattrs(attr, a, &attrs);
	if (err)
		return err;

	/* Metadata attributes. */
	if (attrs & (1 << OVS_KEY_ATTR_PRIORITY)) {
		swkey->phy.priority = nla_get_u32(a[OVS_KEY_ATTR_PRIORITY]);
		attrs &= ~(1 << OVS_KEY_ATTR_PRIORITY);
	}
	if (attrs & (1 << OVS_KEY_ATTR_IN_PORT)) {
		u32 in_port = nla_get_u32(a[OVS_KEY_ATTR_IN_PORT]);
		if (in_port >= DP_MAX_PORTS)
			return -EINVAL;
		swkey->phy.in_port = in_port;
		attrs &= ~(1 << OVS_KEY_ATTR_IN_PORT);
	} else {
		swkey->phy.in_port = DP_MAX_PORTS;
	}
	if (attrs & (1 << OVS_KEY_ATTR_SKB_MARK)) {
		swkey->phy.skb_mark = nla_get_u32(a[OVS_KEY_ATTR_SKB_MARK]);
		attrs &= ~(1 << OVS_KEY_ATTR_SKB_MARK);
	}

	if (attrs & (1 << OVS_KEY_ATTR_TUNNEL)) {
		err = ovs_ipv4_tun_from_nlattr(a[OVS_KEY_ATTR_TUNNEL], &swkey->tun_key);
		if (err)
			return err;

		attrs &= ~(1 << OVS_KEY_ATTR_TUNNEL);
	}

	/* Data attributes. */
	if (!(attrs & (1 << OVS_KEY_ATTR_ETHERNET)))
		return -EINVAL;
	attrs &= ~(1 << OVS_KEY_ATTR_ETHERNET);

	eth_key = nla_data(a[OVS_KEY_ATTR_ETHERNET]);
	memcpy(swkey->eth.src, eth_key->eth_src, ETH_ALEN);
	memcpy(swkey->eth.dst, eth_key->eth_dst, ETH_ALEN);

	if (attrs & (1u << OVS_KEY_ATTR_ETHERTYPE) &&
	    nla_get_be16(a[OVS_KEY_ATTR_ETHERTYPE]) == htons(ETH_P_8021Q)) {
		const struct nlattr *encap;
		__be16 tci;

		if (attrs != ((1 << OVS_KEY_ATTR_VLAN) |
			      (1 << OVS_KEY_ATTR_ETHERTYPE) |
			      (1 << OVS_KEY_ATTR_ENCAP)))
			return -EINVAL;

		encap = a[OVS_KEY_ATTR_ENCAP];
		tci = nla_get_be16(a[OVS_KEY_ATTR_VLAN]);
		if (tci & htons(VLAN_TAG_PRESENT)) {
			swkey->eth.tci = tci;

			err = parse_flow_nlattrs(encap, a, &attrs);
			if (err)
				return err;
		} else if (!tci) {
			/* Corner case for truncated 802.1Q header. */
			if (nla_len(encap))
				return -EINVAL;

			swkey->eth.type = htons(ETH_P_8021Q);
			*key_lenp = key_len;
			return 0;
		} else {
			return -EINVAL;
		}
	}

	if (attrs & (1 << OVS_KEY_ATTR_ETHERTYPE)) {
		swkey->eth.type = nla_get_be16(a[OVS_KEY_ATTR_ETHERTYPE]);
		if (ntohs(swkey->eth.type) < ETH_P_802_3_MIN)
			return -EINVAL;
		attrs &= ~(1 << OVS_KEY_ATTR_ETHERTYPE);
	} else {
		swkey->eth.type = htons(ETH_P_802_2);
	}

	if (swkey->eth.type == htons(ETH_P_IP)) {
		const struct ovs_key_ipv4 *ipv4_key;

		if (!(attrs & (1 << OVS_KEY_ATTR_IPV4)))
			return -EINVAL;
		attrs &= ~(1 << OVS_KEY_ATTR_IPV4);

		key_len = SW_FLOW_KEY_OFFSET(ipv4.addr);
		ipv4_key = nla_data(a[OVS_KEY_ATTR_IPV4]);
		if (ipv4_key->ipv4_frag > OVS_FRAG_TYPE_MAX)
			return -EINVAL;
		swkey->ip.proto = ipv4_key->ipv4_proto;
		swkey->ip.tos = ipv4_key->ipv4_tos;
		swkey->ip.ttl = ipv4_key->ipv4_ttl;
		swkey->ip.frag = ipv4_key->ipv4_frag;
		swkey->ipv4.addr.src = ipv4_key->ipv4_src;
		swkey->ipv4.addr.dst = ipv4_key->ipv4_dst;

		if (swkey->ip.frag != OVS_FRAG_TYPE_LATER) {
			err = ipv4_flow_from_nlattrs(swkey, &key_len, a, &attrs);
			if (err)
				return err;
		}
	} else if (swkey->eth.type == htons(ETH_P_IPV6)) {
		const struct ovs_key_ipv6 *ipv6_key;

		if (!(attrs & (1 << OVS_KEY_ATTR_IPV6)))
			return -EINVAL;
		attrs &= ~(1 << OVS_KEY_ATTR_IPV6);

		key_len = SW_FLOW_KEY_OFFSET(ipv6.label);
		ipv6_key = nla_data(a[OVS_KEY_ATTR_IPV6]);
		if (ipv6_key->ipv6_frag > OVS_FRAG_TYPE_MAX)
			return -EINVAL;
		swkey->ipv6.label = ipv6_key->ipv6_label;
		swkey->ip.proto = ipv6_key->ipv6_proto;
		swkey->ip.tos = ipv6_key->ipv6_tclass;
		swkey->ip.ttl = ipv6_key->ipv6_hlimit;
		swkey->ip.frag = ipv6_key->ipv6_frag;
		memcpy(&swkey->ipv6.addr.src, ipv6_key->ipv6_src,
		       sizeof(swkey->ipv6.addr.src));
		memcpy(&swkey->ipv6.addr.dst, ipv6_key->ipv6_dst,
		       sizeof(swkey->ipv6.addr.dst));

		if (swkey->ip.frag != OVS_FRAG_TYPE_LATER) {
			err = ipv6_flow_from_nlattrs(swkey, &key_len, a, &attrs);
			if (err)
				return err;
		}
	} else if (swkey->eth.type == htons(ETH_P_ARP) ||
		   swkey->eth.type == htons(ETH_P_RARP)) {
		const struct ovs_key_arp *arp_key;

		if (!(attrs & (1 << OVS_KEY_ATTR_ARP)))
			return -EINVAL;
		attrs &= ~(1 << OVS_KEY_ATTR_ARP);

		key_len = SW_FLOW_KEY_OFFSET(ipv4.arp);
		arp_key = nla_data(a[OVS_KEY_ATTR_ARP]);
		swkey->ipv4.addr.src = arp_key->arp_sip;
		swkey->ipv4.addr.dst = arp_key->arp_tip;
		if (arp_key->arp_op & htons(0xff00))
			return -EINVAL;
		swkey->ip.proto = ntohs(arp_key->arp_op);
		memcpy(swkey->ipv4.arp.sha, arp_key->arp_sha, ETH_ALEN);
		memcpy(swkey->ipv4.arp.tha, arp_key->arp_tha, ETH_ALEN);
	}

	if (attrs)
		return -EINVAL;
	*key_lenp = key_len;

	return 0;
}

/**
 * ovs_flow_metadata_from_nlattrs - parses Netlink attributes into a flow key.
 * @flow: Receives extracted in_port, priority, tun_key and skb_mark.
 * @key_len: Length of key in @flow.  Used for calculating flow hash.
 * @attr: Netlink attribute holding nested %OVS_KEY_ATTR_* Netlink attribute
 * sequence.
 *
 * This parses a series of Netlink attributes that form a flow key, which must
 * take the same form accepted by flow_from_nlattrs(), but only enough of it to
 * get the metadata, that is, the parts of the flow key that cannot be
 * extracted from the packet itself.
 */
int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow, int key_len,
				   const struct nlattr *attr)
{
	struct ovs_key_ipv4_tunnel *tun_key = &flow->key.tun_key;
	const struct nlattr *nla;
	int rem;

	flow->key.phy.in_port = DP_MAX_PORTS;
	flow->key.phy.priority = 0;
	flow->key.phy.skb_mark = 0;
	memset(tun_key, 0, sizeof(flow->key.tun_key));

	nla_for_each_nested(nla, attr, rem) {
		int type = nla_type(nla);

		if (type <= OVS_KEY_ATTR_MAX && ovs_key_lens[type] > 0) {
			int err;

			if (nla_len(nla) != ovs_key_lens[type])
				return -EINVAL;

			switch (type) {
			case OVS_KEY_ATTR_PRIORITY:
				flow->key.phy.priority = nla_get_u32(nla);
				break;

			case OVS_KEY_ATTR_TUNNEL:
				err = ovs_ipv4_tun_from_nlattr(nla, tun_key);
				if (err)
					return err;
				break;

			case OVS_KEY_ATTR_IN_PORT:
				if (nla_get_u32(nla) >= DP_MAX_PORTS)
					return -EINVAL;
				flow->key.phy.in_port = nla_get_u32(nla);
				break;

			case OVS_KEY_ATTR_SKB_MARK:
				flow->key.phy.skb_mark = nla_get_u32(nla);
				break;
			}
		}
	}
	if (rem)
		return -EINVAL;

	flow->hash = ovs_flow_hash(&flow->key,
				   flow_key_start(&flow->key), key_len);

	return 0;
}

int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb)
{
	struct ovs_key_ethernet *eth_key;
	struct nlattr *nla, *encap;

	if (swkey->phy.priority &&
	    nla_put_u32(skb, OVS_KEY_ATTR_PRIORITY, swkey->phy.priority))
		goto nla_put_failure;

	if (swkey->tun_key.ipv4_dst &&
	    ovs_ipv4_tun_to_nlattr(skb, &swkey->tun_key))
		goto nla_put_failure;

	if (swkey->phy.in_port != DP_MAX_PORTS &&
	    nla_put_u32(skb, OVS_KEY_ATTR_IN_PORT, swkey->phy.in_port))
		goto nla_put_failure;

	if (swkey->phy.skb_mark &&
	    nla_put_u32(skb, OVS_KEY_ATTR_SKB_MARK, swkey->phy.skb_mark))
		goto nla_put_failure;

	nla = nla_reserve(skb, OVS_KEY_ATTR_ETHERNET, sizeof(*eth_key));
	if (!nla)
		goto nla_put_failure;
	eth_key = nla_data(nla);
	memcpy(eth_key->eth_src, swkey->eth.src, ETH_ALEN);
	memcpy(eth_key->eth_dst, swkey->eth.dst, ETH_ALEN);

	if (swkey->eth.tci || swkey->eth.type == htons(ETH_P_8021Q)) {
		if (nla_put_be16(skb, OVS_KEY_ATTR_ETHERTYPE, htons(ETH_P_8021Q)) ||
		    nla_put_be16(skb, OVS_KEY_ATTR_VLAN, swkey->eth.tci))
			goto nla_put_failure;
		encap = nla_nest_start(skb, OVS_KEY_ATTR_ENCAP);
		if (!swkey->eth.tci)
			goto unencap;
	} else {
		encap = NULL;
	}

	if (swkey->eth.type == htons(ETH_P_802_2))
		goto unencap;

	if (nla_put_be16(skb, OVS_KEY_ATTR_ETHERTYPE, swkey->eth.type))
		goto nla_put_failure;

	if (swkey->eth.type == htons(ETH_P_IP)) {
		struct ovs_key_ipv4 *ipv4_key;

		nla = nla_reserve(skb, OVS_KEY_ATTR_IPV4, sizeof(*ipv4_key));
		if (!nla)
			goto nla_put_failure;
		ipv4_key = nla_data(nla);
		ipv4_key->ipv4_src = swkey->ipv4.addr.src;
		ipv4_key->ipv4_dst = swkey->ipv4.addr.dst;
		ipv4_key->ipv4_proto = swkey->ip.proto;
		ipv4_key->ipv4_tos = swkey->ip.tos;
		ipv4_key->ipv4_ttl = swkey->ip.ttl;
		ipv4_key->ipv4_frag = swkey->ip.frag;
	} else if (swkey->eth.type == htons(ETH_P_IPV6)) {
		struct ovs_key_ipv6 *ipv6_key;

		nla = nla_reserve(skb, OVS_KEY_ATTR_IPV6, sizeof(*ipv6_key));
		if (!nla)
			goto nla_put_failure;
		ipv6_key = nla_data(nla);
		memcpy(ipv6_key->ipv6_src, &swkey->ipv6.addr.src,
				sizeof(ipv6_key->ipv6_src));
		memcpy(ipv6_key->ipv6_dst, &swkey->ipv6.addr.dst,
				sizeof(ipv6_key->ipv6_dst));
		ipv6_key->ipv6_label = swkey->ipv6.label;
		ipv6_key->ipv6_proto = swkey->ip.proto;
		ipv6_key->ipv6_tclass = swkey->ip.tos;
		ipv6_key->ipv6_hlimit = swkey->ip.ttl;
		ipv6_key->ipv6_frag = swkey->ip.frag;
	} else if (swkey->eth.type == htons(ETH_P_ARP) ||
		   swkey->eth.type == htons(ETH_P_RARP)) {
		struct ovs_key_arp *arp_key;

		nla = nla_reserve(skb, OVS_KEY_ATTR_ARP, sizeof(*arp_key));
		if (!nla)
			goto nla_put_failure;
		arp_key = nla_data(nla);
		memset(arp_key, 0, sizeof(struct ovs_key_arp));
		arp_key->arp_sip = swkey->ipv4.addr.src;
		arp_key->arp_tip = swkey->ipv4.addr.dst;
		arp_key->arp_op = htons(swkey->ip.proto);
		memcpy(arp_key->arp_sha, swkey->ipv4.arp.sha, ETH_ALEN);
		memcpy(arp_key->arp_tha, swkey->ipv4.arp.tha, ETH_ALEN);
	}

	if ((swkey->eth.type == htons(ETH_P_IP) ||
	     swkey->eth.type == htons(ETH_P_IPV6)) &&
	     swkey->ip.frag != OVS_FRAG_TYPE_LATER) {

		if (swkey->ip.proto == IPPROTO_TCP) {
			struct ovs_key_tcp *tcp_key;

			nla = nla_reserve(skb, OVS_KEY_ATTR_TCP, sizeof(*tcp_key));
			if (!nla)
				goto nla_put_failure;
			tcp_key = nla_data(nla);
			if (swkey->eth.type == htons(ETH_P_IP)) {
				tcp_key->tcp_src = swkey->ipv4.tp.src;
				tcp_key->tcp_dst = swkey->ipv4.tp.dst;
			} else if (swkey->eth.type == htons(ETH_P_IPV6)) {
				tcp_key->tcp_src = swkey->ipv6.tp.src;
				tcp_key->tcp_dst = swkey->ipv6.tp.dst;
			}
		} else if (swkey->ip.proto == IPPROTO_UDP) {
			struct ovs_key_udp *udp_key;

			nla = nla_reserve(skb, OVS_KEY_ATTR_UDP, sizeof(*udp_key));
			if (!nla)
				goto nla_put_failure;
			udp_key = nla_data(nla);
			if (swkey->eth.type == htons(ETH_P_IP)) {
				udp_key->udp_src = swkey->ipv4.tp.src;
				udp_key->udp_dst = swkey->ipv4.tp.dst;
			} else if (swkey->eth.type == htons(ETH_P_IPV6)) {
				udp_key->udp_src = swkey->ipv6.tp.src;
				udp_key->udp_dst = swkey->ipv6.tp.dst;
			}
		} else if (swkey->eth.type == htons(ETH_P_IP) &&
			   swkey->ip.proto == IPPROTO_ICMP) {
			struct ovs_key_icmp *icmp_key;

			nla = nla_reserve(skb, OVS_KEY_ATTR_ICMP, sizeof(*icmp_key));
			if (!nla)
				goto nla_put_failure;
			icmp_key = nla_data(nla);
			icmp_key->icmp_type = ntohs(swkey->ipv4.tp.src);
			icmp_key->icmp_code = ntohs(swkey->ipv4.tp.dst);
		} else if (swkey->eth.type == htons(ETH_P_IPV6) &&
			   swkey->ip.proto == IPPROTO_ICMPV6) {
			struct ovs_key_icmpv6 *icmpv6_key;

			nla = nla_reserve(skb, OVS_KEY_ATTR_ICMPV6,
						sizeof(*icmpv6_key));
			if (!nla)
				goto nla_put_failure;
			icmpv6_key = nla_data(nla);
			icmpv6_key->icmpv6_type = ntohs(swkey->ipv6.tp.src);
			icmpv6_key->icmpv6_code = ntohs(swkey->ipv6.tp.dst);

			if (icmpv6_key->icmpv6_type == NDISC_NEIGHBOUR_SOLICITATION ||
			    icmpv6_key->icmpv6_type == NDISC_NEIGHBOUR_ADVERTISEMENT) {
				struct ovs_key_nd *nd_key;

				nla = nla_reserve(skb, OVS_KEY_ATTR_ND, sizeof(*nd_key));
				if (!nla)
					goto nla_put_failure;
				nd_key = nla_data(nla);
				memcpy(nd_key->nd_target, &swkey->ipv6.nd.target,
							sizeof(nd_key->nd_target));
				memcpy(nd_key->nd_sll, swkey->ipv6.nd.sll, ETH_ALEN);
				memcpy(nd_key->nd_tll, swkey->ipv6.nd.tll, ETH_ALEN);
			}
		}
	}

unencap:
	if (encap)
		nla_nest_end(skb, encap);

	return 0;

nla_put_failure:
	return -EMSGSIZE;
}

/* Initializes the flow module.
 * Returns zero if successful or a negative error code. */
int ovs_flow_init(void)
{
	flow_cache = kmem_cache_create("sw_flow", sizeof(struct sw_flow), 0,
					0, NULL);
	if (flow_cache == NULL)
		return -ENOMEM;

	return 0;
}

/* Uninitializes the flow module. */
void ovs_flow_exit(void)
{
	kmem_cache_destroy(flow_cache);
}