/*
- * Copyright (c) 2009, 2010, 2011, 2012, 2013, 2014, 2015 Nicira, Inc.
+ * Copyright (c) 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016 Nicira, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
#include "csum.h"
#include "crc32c.h"
#include "flow.h"
-#include "hmap.h"
-#include "dynamic-string.h"
+#include "openvswitch/hmap.h"
+#include "openvswitch/dynamic-string.h"
#include "ovs-thread.h"
#include "odp-util.h"
#include "dp-packet.h"
const struct in6_addr in6addr_exact = IN6ADDR_EXACT_INIT;
const struct in6_addr in6addr_all_hosts = IN6ADDR_ALL_HOSTS_INIT;
+const struct in6_addr in6addr_all_routers = IN6ADDR_ALL_ROUTERS_INIT;
struct in6_addr
flow_tnl_dst(const struct flow_tnl *tnl)
return false;
}
+/* Attempts to parse 's' as an Ethernet address. If successful, stores the
+ * address in 'ea' and returns true, otherwise zeros 'ea' and returns
+ * false. This function checks trailing characters. */
bool
eth_addr_from_string(const char *s, struct eth_addr *ea)
{
- if (ovs_scan(s, ETH_ADDR_SCAN_FMT, ETH_ADDR_SCAN_ARGS(*ea))) {
+ int n = 0;
+ if (ovs_scan(s, ETH_ADDR_SCAN_FMT"%n", ETH_ADDR_SCAN_ARGS(*ea), &n)
+ && !s[n]) {
return true;
} else {
*ea = eth_addr_zero;
dp_packet_reset_offsets(b);
dp_packet_set_l3(b, arp);
+ b->packet_type = htonl(PT_ETH);
}
/* Insert VLAN header according to given TCI. Packet passed must be Ethernet
void
eth_pop_vlan(struct dp_packet *packet)
{
- struct vlan_eth_header *veh = dp_packet_l2(packet);
+ struct vlan_eth_header *veh = dp_packet_eth(packet);
if (veh && dp_packet_size(packet) >= sizeof *veh
&& eth_type_vlan(veh->veth_type)) {
}
}
+/* Push Ethernet header onto 'packet' assuming it is layer 3 */
+void
+push_eth(struct dp_packet *packet, const struct eth_addr *dst,
+ const struct eth_addr *src)
+{
+ struct eth_header *eh;
+
+ ovs_assert(packet->packet_type != htonl(PT_ETH));
+ eh = dp_packet_resize_l2(packet, ETH_HEADER_LEN);
+ eh->eth_dst = *dst;
+ eh->eth_src = *src;
+ eh->eth_type = pt_ns_type_be(packet->packet_type);
+ packet->packet_type = htonl(PT_ETH);
+}
+
+/* Removes Ethernet header, including VLAN header, from 'packet'.
+ *
+ * Previous to calling this function, 'ofpbuf_l3(packet)' must not be NULL */
+void
+pop_eth(struct dp_packet *packet)
+{
+ char *l2_5 = dp_packet_l2_5(packet);
+ char *l3 = dp_packet_l3(packet);
+ ovs_be16 ethertype;
+ int increment;
+
+ ovs_assert(packet->packet_type == htonl(PT_ETH));
+ ovs_assert(l3 != NULL);
+
+ if (l2_5) {
+ increment = packet->l2_5_ofs;
+ ethertype = *(ALIGNED_CAST(ovs_be16 *, (l2_5 - 2)));
+ } else {
+ increment = packet->l3_ofs;
+ ethertype = *(ALIGNED_CAST(ovs_be16 *, (l3 - 2)));
+ }
+
+ dp_packet_resize_l2(packet, -increment);
+ packet->packet_type = PACKET_TYPE_BE(OFPHTN_ETHERTYPE, ntohs(ethertype));
+}
+
/* Set ethertype of the packet. */
static void
set_ethertype(struct dp_packet *packet, ovs_be16 eth_type)
{
- struct eth_header *eh = dp_packet_l2(packet);
+ struct eth_header *eh = dp_packet_eth(packet);
if (!eh) {
return;
}
}
+/* Parses string 's', which must be an IP address. Stores the IP address into
+ * '*ip'. Returns true if successful, otherwise false. */
+bool
+ip_parse(const char *s, ovs_be32 *ip)
+{
+ return inet_pton(AF_INET, s, ip) == 1;
+}
+
+/* Parses string 's', which must be an IP address with a port number
+ * with ":" as a separator (e.g.: 192.168.1.2:80).
+ * Stores the IP address into '*ip' and port number to '*port'.
+ *
+ * Returns NULL if successful, otherwise an error message that the caller must
+ * free(). */
+char * OVS_WARN_UNUSED_RESULT
+ip_parse_port(const char *s, ovs_be32 *ip, ovs_be16 *port)
+{
+ int n = 0;
+ if (ovs_scan(s, IP_PORT_SCAN_FMT"%n", IP_PORT_SCAN_ARGS(ip, port), &n)
+ && !s[n]) {
+ return NULL;
+ }
+
+ return xasprintf("%s: invalid IP address or port number", s);
+}
+
/* Parses string 's', which must be an IP address with an optional netmask or
- * CIDR prefix length. Stores the IP address into '*ip' and the netmask into
- * '*mask'. (If 's' does not contain a netmask, 255.255.255.255 is
- * assumed.)
+ * CIDR prefix length. Stores the IP address into '*ip', netmask into '*mask',
+ * (255.255.255.255, if 's' lacks a netmask), and number of scanned characters
+ * into '*n'.
*
* Returns NULL if successful, otherwise an error message that the caller must
* free(). */
char * OVS_WARN_UNUSED_RESULT
-ip_parse_masked(const char *s, ovs_be32 *ip, ovs_be32 *mask)
+ip_parse_masked_len(const char *s, int *n, ovs_be32 *ip,
+ ovs_be32 *mask)
{
int prefix;
- int n;
- if (ovs_scan(s, IP_SCAN_FMT"/"IP_SCAN_FMT"%n",
- IP_SCAN_ARGS(ip), IP_SCAN_ARGS(mask), &n) && !s[n]) {
+ if (ovs_scan_len(s, n, IP_SCAN_FMT"/"IP_SCAN_FMT,
+ IP_SCAN_ARGS(ip), IP_SCAN_ARGS(mask))) {
/* OK. */
- } else if (ovs_scan(s, IP_SCAN_FMT"/%d%n", IP_SCAN_ARGS(ip), &prefix, &n)
- && !s[n]) {
- if (prefix <= 0 || prefix > 32) {
- return xasprintf("%s: network prefix bits not between 0 and "
- "32", s);
+ } else if (ovs_scan_len(s, n, IP_SCAN_FMT"/%d",
+ IP_SCAN_ARGS(ip), &prefix)) {
+ if (prefix < 0 || prefix > 32) {
+ return xasprintf("%s: IPv4 network prefix bits not between 0 and "
+ "32, inclusive", s);
}
*mask = be32_prefix_mask(prefix);
- } else if (ovs_scan(s, IP_SCAN_FMT"%n", IP_SCAN_ARGS(ip), &n) && !s[n]) {
+ } else if (ovs_scan_len(s, n, IP_SCAN_FMT, IP_SCAN_ARGS(ip))) {
*mask = OVS_BE32_MAX;
} else {
return xasprintf("%s: invalid IP address", s);
return NULL;
}
+/* This function is similar to ip_parse_masked_len(), but doesn't return the
+ * number of scanned characters and expects 's' to end after the ip/(optional)
+ * mask.
+ *
+ * Returns NULL if successful, otherwise an error message that the caller must
+ * free(). */
+char * OVS_WARN_UNUSED_RESULT
+ip_parse_masked(const char *s, ovs_be32 *ip, ovs_be32 *mask)
+{
+ int n = 0;
+
+ char *error = ip_parse_masked_len(s, &n, ip, mask);
+ if (!error && s[n]) {
+ return xasprintf("%s: invalid IP address", s);
+ }
+ return error;
+}
+
+/* Similar to ip_parse_masked_len(), but the mask, if present, must be a CIDR
+ * mask and is returned as a prefix len in '*plen'. */
+char * OVS_WARN_UNUSED_RESULT
+ip_parse_cidr_len(const char *s, int *n, ovs_be32 *ip, unsigned int *plen)
+{
+ ovs_be32 mask;
+ char *error;
+
+ error = ip_parse_masked_len(s, n, ip, &mask);
+ if (error) {
+ return error;
+ }
+
+ if (!ip_is_cidr(mask)) {
+ return xasprintf("%s: CIDR network required", s);
+ }
+ *plen = ip_count_cidr_bits(mask);
+ return NULL;
+}
+
+/* Similar to ip_parse_cidr_len(), but doesn't return the number of scanned
+ * characters and expects 's' to be NULL terminated at the end of the
+ * ip/(optional) cidr. */
+char * OVS_WARN_UNUSED_RESULT
+ip_parse_cidr(const char *s, ovs_be32 *ip, unsigned int *plen)
+{
+ int n = 0;
+
+ char *error = ip_parse_cidr_len(s, &n, ip, plen);
+ if (!error && s[n]) {
+ return xasprintf("%s: invalid IP address", s);
+ }
+ return error;
+}
+
+/* Parses string 's', which must be an IPv6 address. Stores the IPv6 address
+ * into '*ip'. Returns true if successful, otherwise false. */
+bool
+ipv6_parse(const char *s, struct in6_addr *ip)
+{
+ return inet_pton(AF_INET6, s, ip) == 1;
+}
+
+/* Parses string 's', which must be an IPv6 address with an optional netmask or
+ * CIDR prefix length. Stores the IPv6 address into '*ip' and the netmask into
+ * '*mask' (if 's' does not contain a netmask, all-one-bits is assumed), and
+ * number of scanned characters into '*n'.
+ *
+ * Returns NULL if successful, otherwise an error message that the caller must
+ * free(). */
+char * OVS_WARN_UNUSED_RESULT
+ipv6_parse_masked_len(const char *s, int *n, struct in6_addr *ip,
+ struct in6_addr *mask)
+{
+ char ipv6_s[IPV6_SCAN_LEN + 1];
+ int prefix;
+
+ if (ovs_scan_len(s, n, " "IPV6_SCAN_FMT, ipv6_s)
+ && ipv6_parse(ipv6_s, ip)) {
+ if (ovs_scan_len(s, n, "/%d", &prefix)) {
+ if (prefix < 0 || prefix > 128) {
+ return xasprintf("%s: IPv6 network prefix bits not between 0 "
+ "and 128, inclusive", s);
+ }
+ *mask = ipv6_create_mask(prefix);
+ } else if (ovs_scan_len(s, n, "/"IPV6_SCAN_FMT, ipv6_s)) {
+ if (!ipv6_parse(ipv6_s, mask)) {
+ return xasprintf("%s: Invalid IPv6 mask", s);
+ }
+ /* OK. */
+ } else {
+ /* OK. No mask. */
+ *mask = in6addr_exact;
+ }
+ return NULL;
+ }
+ return xasprintf("%s: invalid IPv6 address", s);
+}
+
+/* This function is similar to ipv6_parse_masked_len(), but doesn't return the
+ * number of scanned characters and expects 's' to end following the
+ * ipv6/(optional) mask. */
+char * OVS_WARN_UNUSED_RESULT
+ipv6_parse_masked(const char *s, struct in6_addr *ip, struct in6_addr *mask)
+{
+ int n = 0;
+
+ char *error = ipv6_parse_masked_len(s, &n, ip, mask);
+ if (!error && s[n]) {
+ return xasprintf("%s: invalid IPv6 address", s);
+ }
+ return error;
+}
+
+/* Similar to ipv6_parse_masked_len(), but the mask, if present, must be a CIDR
+ * mask and is returned as a prefix length in '*plen'. */
+char * OVS_WARN_UNUSED_RESULT
+ipv6_parse_cidr_len(const char *s, int *n, struct in6_addr *ip,
+ unsigned int *plen)
+{
+ struct in6_addr mask;
+ char *error;
+
+ error = ipv6_parse_masked_len(s, n, ip, &mask);
+ if (error) {
+ return error;
+ }
+
+ if (!ipv6_is_cidr(&mask)) {
+ return xasprintf("%s: IPv6 CIDR network required", s);
+ }
+ *plen = ipv6_count_cidr_bits(&mask);
+ return NULL;
+}
+
+/* Similar to ipv6_parse_cidr_len(), but doesn't return the number of scanned
+ * characters and expects 's' to end after the ipv6/(optional) cidr. */
+char * OVS_WARN_UNUSED_RESULT
+ipv6_parse_cidr(const char *s, struct in6_addr *ip, unsigned int *plen)
+{
+ int n = 0;
+
+ char *error = ipv6_parse_cidr_len(s, &n, ip, plen);
+ if (!error && s[n]) {
+ return xasprintf("%s: invalid IPv6 address", s);
+ }
+ return error;
+}
+
+/* Stores the string representation of the IPv6 address 'addr' into the
+ * character array 'addr_str', which must be at least INET6_ADDRSTRLEN
+ * bytes long. */
void
ipv6_format_addr(const struct in6_addr *addr, struct ds *s)
{
}
}
-struct in6_addr ipv6_addr_bitand(const struct in6_addr *a,
- const struct in6_addr *b)
-{
- int i;
- struct in6_addr dst;
-
#ifdef s6_addr32
- for (i=0; i<4; i++) {
- dst.s6_addr32[i] = a->s6_addr32[i] & b->s6_addr32[i];
- }
+#define s6_addrX s6_addr32
+#define IPV6_FOR_EACH(VAR) for (int VAR = 0; VAR < 4; VAR++)
#else
- for (i=0; i<16; i++) {
- dst.s6_addr[i] = a->s6_addr[i] & b->s6_addr[i];
- }
+#define s6_addrX s6_addr
+#define IPV6_FOR_EACH(VAR) for (int VAR = 0; VAR < 16; VAR++)
#endif
- return dst;
+struct in6_addr
+ipv6_addr_bitand(const struct in6_addr *a, const struct in6_addr *b)
+{
+ struct in6_addr dst;
+ IPV6_FOR_EACH (i) {
+ dst.s6_addrX[i] = a->s6_addrX[i] & b->s6_addrX[i];
+ }
+ return dst;
+}
+
+struct in6_addr
+ipv6_addr_bitxor(const struct in6_addr *a, const struct in6_addr *b)
+{
+ struct in6_addr dst;
+ IPV6_FOR_EACH (i) {
+ dst.s6_addrX[i] = a->s6_addrX[i] ^ b->s6_addrX[i];
+ }
+ return dst;
+}
+
+bool
+ipv6_is_zero(const struct in6_addr *a)
+{
+ IPV6_FOR_EACH (i) {
+ if (a->s6_addrX[i]) {
+ return false;
+ }
+ }
+ return true;
}
/* Returns an in6_addr consisting of 'mask' high-order 1-bits and 128-N
return true;
}
-/* Parses string 's', which must be an IPv6 address with an optional
- * CIDR prefix length. Stores the IP address into '*ipv6' and the CIDR
- * prefix in '*prefix'. (If 's' does not contain a CIDR length, all-ones
- * is assumed.)
- *
- * Returns NULL if successful, otherwise an error message that the caller must
- * free(). */
-char * OVS_WARN_UNUSED_RESULT
-ipv6_parse_masked(const char *s, struct in6_addr *ipv6, struct in6_addr *mask)
-{
- char ipv6_s[IPV6_SCAN_LEN + 1];
- char mask_s[IPV6_SCAN_LEN + 1];
- int prefix;
- int n;
-
- if (ovs_scan(s, IPV6_SCAN_FMT"/"IPV6_SCAN_FMT"%n", ipv6_s, mask_s, &n)
- && inet_pton(AF_INET6, ipv6_s, ipv6) == 1
- && inet_pton(AF_INET6, mask_s, mask) == 1
- && !s[n]) {
- /* OK. */
- } else if (ovs_scan(s, IPV6_SCAN_FMT"/%d%n", ipv6_s, &prefix, &n)
- && inet_pton(AF_INET6, ipv6_s, ipv6) == 1
- && !s[n]) {
- if (prefix <= 0 || prefix > 128) {
- return xasprintf("%s: prefix bits not between 0 and 128", s);
- }
- *mask = ipv6_create_mask(prefix);
- } else if (ovs_scan(s, IPV6_SCAN_FMT"%n", ipv6_s, &n)
- && inet_pton(AF_INET6, ipv6_s, ipv6) == 1
- && !s[n]) {
- *mask = in6addr_exact;
- } else {
- return xasprintf("%s: invalid IP address", s);
- }
- return NULL;
-}
-
/* Populates 'b' with an Ethernet II packet headed with the given 'eth_dst',
* 'eth_src' and 'eth_type' parameters. A payload of 'size' bytes is allocated
* in 'b' and returned. This payload may be populated with appropriate
dp_packet_prealloc_tailroom(b, 2 + ETH_HEADER_LEN + VLAN_HEADER_LEN + size);
dp_packet_reserve(b, 2 + VLAN_HEADER_LEN);
eth = dp_packet_put_uninit(b, ETH_HEADER_LEN);
- data = dp_packet_put_uninit(b, size);
+ data = dp_packet_put_zeros(b, size);
eth->eth_dst = eth_dst;
eth->eth_src = eth_src;
eth->eth_type = htons(eth_type);
+ b->packet_type = htonl(PT_ETH);
dp_packet_reset_offsets(b);
dp_packet_set_l3(b, data);
return data;
}
-static void
+void
packet_set_ipv4_addr(struct dp_packet *packet,
ovs_16aligned_be32 *addr, ovs_be32 new_addr)
{
*
* This function assumes that L3 and L4 offsets are set in the packet. */
static bool
-packet_rh_present(struct dp_packet *packet)
+packet_rh_present(struct dp_packet *packet, uint8_t *nexthdr)
{
const struct ovs_16aligned_ip6_hdr *nh;
- int nexthdr;
size_t len;
size_t remaining;
uint8_t *data = dp_packet_l3(packet);
remaining = packet->l4_ofs - packet->l3_ofs;
-
if (remaining < sizeof *nh) {
return false;
}
nh = ALIGNED_CAST(struct ovs_16aligned_ip6_hdr *, data);
data += sizeof *nh;
remaining -= sizeof *nh;
- nexthdr = nh->ip6_nxt;
+ *nexthdr = nh->ip6_nxt;
while (1) {
- if ((nexthdr != IPPROTO_HOPOPTS)
- && (nexthdr != IPPROTO_ROUTING)
- && (nexthdr != IPPROTO_DSTOPTS)
- && (nexthdr != IPPROTO_AH)
- && (nexthdr != IPPROTO_FRAGMENT)) {
+ if ((*nexthdr != IPPROTO_HOPOPTS)
+ && (*nexthdr != IPPROTO_ROUTING)
+ && (*nexthdr != IPPROTO_DSTOPTS)
+ && (*nexthdr != IPPROTO_AH)
+ && (*nexthdr != IPPROTO_FRAGMENT)) {
/* It's either a terminal header (e.g., TCP, UDP) or one we
* don't understand. In either case, we're done with the
* packet, so use it to fill in 'nw_proto'. */
return false;
}
- if (nexthdr == IPPROTO_AH) {
+ if (*nexthdr == IPPROTO_AH) {
/* A standard AH definition isn't available, but the fields
* we care about are in the same location as the generic
* option header--only the header length is calculated
* differently. */
const struct ip6_ext *ext_hdr = (struct ip6_ext *)data;
- nexthdr = ext_hdr->ip6e_nxt;
+ *nexthdr = ext_hdr->ip6e_nxt;
len = (ext_hdr->ip6e_len + 2) * 4;
- } else if (nexthdr == IPPROTO_FRAGMENT) {
+ } else if (*nexthdr == IPPROTO_FRAGMENT) {
const struct ovs_16aligned_ip6_frag *frag_hdr
= ALIGNED_CAST(struct ovs_16aligned_ip6_frag *, data);
- nexthdr = frag_hdr->ip6f_nxt;
+ *nexthdr = frag_hdr->ip6f_nxt;
len = sizeof *frag_hdr;
- } else if (nexthdr == IPPROTO_ROUTING) {
+ } else if (*nexthdr == IPPROTO_ROUTING) {
const struct ip6_rthdr *rh = (struct ip6_rthdr *)data;
if (rh->ip6r_segleft > 0) {
return true;
}
- nexthdr = rh->ip6r_nxt;
+ *nexthdr = rh->ip6r_nxt;
len = (rh->ip6r_len + 1) * 8;
} else {
const struct ip6_ext *ext_hdr = (struct ip6_ext *)data;
- nexthdr = ext_hdr->ip6e_nxt;
+ *nexthdr = ext_hdr->ip6e_nxt;
len = (ext_hdr->ip6e_len + 1) * 8;
}
static void
packet_update_csum128(struct dp_packet *packet, uint8_t proto,
- ovs_16aligned_be32 addr[4], const ovs_be32 new_addr[4])
+ ovs_16aligned_be32 addr[4],
+ const struct in6_addr *new_addr)
{
size_t l4_size = dp_packet_l4_size(packet);
}
}
-static void
+void
packet_set_ipv6_addr(struct dp_packet *packet, uint8_t proto,
- ovs_16aligned_be32 addr[4], const ovs_be32 new_addr[4],
+ ovs_16aligned_be32 addr[4],
+ const struct in6_addr *new_addr,
bool recalculate_csum)
{
if (recalculate_csum) {
* appropriate. 'packet' must contain a valid IPv6 packet with correctly
* populated l[34] offsets. */
void
-packet_set_ipv6(struct dp_packet *packet, uint8_t proto, const ovs_be32 src[4],
- const ovs_be32 dst[4], uint8_t key_tc, ovs_be32 key_fl,
+packet_set_ipv6(struct dp_packet *packet, const struct in6_addr *src,
+ const struct in6_addr *dst, uint8_t key_tc, ovs_be32 key_fl,
uint8_t key_hl)
{
struct ovs_16aligned_ip6_hdr *nh = dp_packet_l3(packet);
+ uint8_t proto = 0;
+ bool rh_present;
+
+ rh_present = packet_rh_present(packet, &proto);
if (memcmp(&nh->ip6_src, src, sizeof(ovs_be32[4]))) {
packet_set_ipv6_addr(packet, proto, nh->ip6_src.be32, src, true);
if (memcmp(&nh->ip6_dst, dst, sizeof(ovs_be32[4]))) {
packet_set_ipv6_addr(packet, proto, nh->ip6_dst.be32, dst,
- !packet_rh_present(packet));
+ !rh_present);
}
packet_set_ipv6_tc(&nh->ip6_flow, key_tc);
-
packet_set_ipv6_flow_label(&nh->ip6_flow, key_fl);
-
nh->ip6_hlim = key_hl;
}
}
void
-packet_set_nd(struct dp_packet *packet, const ovs_be32 target[4],
- const struct eth_addr sll, const struct eth_addr tll) {
+packet_set_nd(struct dp_packet *packet, const struct in6_addr *target,
+ const struct eth_addr sll, const struct eth_addr tll)
+{
struct ovs_nd_msg *ns;
- struct ovs_nd_opt *nd_opt;
+ struct ovs_nd_lla_opt *opt;
int bytes_remain = dp_packet_l4_size(packet);
if (OVS_UNLIKELY(bytes_remain < sizeof(*ns))) {
}
ns = dp_packet_l4(packet);
- nd_opt = &ns->options[0];
+ opt = &ns->options[0];
bytes_remain -= sizeof(*ns);
if (memcmp(&ns->target, target, sizeof(ovs_be32[4]))) {
- packet_set_ipv6_addr(packet, IPPROTO_ICMPV6,
- ns->target.be32,
- target, true);
+ packet_set_ipv6_addr(packet, IPPROTO_ICMPV6, ns->target.be32, target,
+ true);
}
- while (bytes_remain >= ND_OPT_LEN && nd_opt->nd_opt_len != 0) {
- if (nd_opt->nd_opt_type == ND_OPT_SOURCE_LINKADDR
- && nd_opt->nd_opt_len == 1) {
- if (!eth_addr_equals(nd_opt->nd_opt_mac, sll)) {
+ while (bytes_remain >= ND_LLA_OPT_LEN && opt->len != 0) {
+ if (opt->type == ND_OPT_SOURCE_LINKADDR && opt->len == 1) {
+ if (!eth_addr_equals(opt->mac, sll)) {
ovs_be16 *csum = &(ns->icmph.icmp6_cksum);
- *csum = recalc_csum48(*csum, nd_opt->nd_opt_mac, sll);
- nd_opt->nd_opt_mac = sll;
+ *csum = recalc_csum48(*csum, opt->mac, sll);
+ opt->mac = sll;
}
/* A packet can only contain one SLL or TLL option */
break;
- } else if (nd_opt->nd_opt_type == ND_OPT_TARGET_LINKADDR
- && nd_opt->nd_opt_len == 1) {
- if (!eth_addr_equals(nd_opt->nd_opt_mac, tll)) {
+ } else if (opt->type == ND_OPT_TARGET_LINKADDR && opt->len == 1) {
+ if (!eth_addr_equals(opt->mac, tll)) {
ovs_be16 *csum = &(ns->icmph.icmp6_cksum);
- *csum = recalc_csum48(*csum, nd_opt->nd_opt_mac, tll);
- nd_opt->nd_opt_mac = tll;
+ *csum = recalc_csum48(*csum, opt->mac, tll);
+ opt->mac = tll;
}
/* A packet can only contain one SLL or TLL option */
break;
}
- nd_opt += nd_opt->nd_opt_len;
- bytes_remain -= nd_opt->nd_opt_len * ND_OPT_LEN;
+ opt += opt->len;
+ bytes_remain -= opt->len * ND_LLA_OPT_LEN;
}
}
* 'arp_op', 'arp_sha', 'arp_tha', 'arp_spa', and 'arp_tpa'. The outer
* Ethernet frame is initialized with Ethernet source 'arp_sha' and destination
* 'arp_tha', except that destination ff:ff:ff:ff:ff:ff is used instead if
- * 'broadcast' is true. */
+ * 'broadcast' is true. Points the L3 header to the ARP header. */
void
compose_arp(struct dp_packet *b, uint16_t arp_op,
const struct eth_addr arp_sha, const struct eth_addr arp_tha,
bool broadcast, ovs_be32 arp_spa, ovs_be32 arp_tpa)
{
- struct eth_header *eth;
- struct arp_eth_header *arp;
+ compose_arp__(b);
+
+ struct eth_header *eth = dp_packet_eth(b);
+ eth->eth_dst = broadcast ? eth_addr_broadcast : arp_tha;
+ eth->eth_src = arp_sha;
+
+ struct arp_eth_header *arp = dp_packet_l3(b);
+ arp->ar_op = htons(arp_op);
+ arp->ar_sha = arp_sha;
+ arp->ar_tha = arp_tha;
+ put_16aligned_be32(&arp->ar_spa, arp_spa);
+ put_16aligned_be32(&arp->ar_tpa, arp_tpa);
+}
+/* Clears 'b' and replaces its contents by an ARP frame. Sets the fields in
+ * the Ethernet and ARP headers that are fixed for ARP frames to those fixed
+ * values, and zeroes the other fields. Points the L3 header to the ARP
+ * header. */
+void
+compose_arp__(struct dp_packet *b)
+{
dp_packet_clear(b);
dp_packet_prealloc_tailroom(b, ARP_PACKET_SIZE);
dp_packet_reserve(b, 2 + VLAN_HEADER_LEN);
- eth = dp_packet_put_uninit(b, sizeof *eth);
- eth->eth_dst = broadcast ? eth_addr_broadcast : arp_tha;
- eth->eth_src = arp_sha;
+ struct eth_header *eth = dp_packet_put_zeros(b, sizeof *eth);
eth->eth_type = htons(ETH_TYPE_ARP);
- arp = dp_packet_put_uninit(b, sizeof *arp);
+ struct arp_eth_header *arp = dp_packet_put_zeros(b, sizeof *arp);
arp->ar_hrd = htons(ARP_HRD_ETHERNET);
arp->ar_pro = htons(ARP_PRO_IP);
arp->ar_hln = sizeof arp->ar_sha;
arp->ar_pln = sizeof arp->ar_spa;
- arp->ar_op = htons(arp_op);
- arp->ar_sha = arp_sha;
- arp->ar_tha = arp_tha;
-
- put_16aligned_be32(&arp->ar_spa, arp_spa);
- put_16aligned_be32(&arp->ar_tpa, arp_tpa);
dp_packet_reset_offsets(b);
dp_packet_set_l3(b, arp);
+
+ b->packet_type = htonl(PT_ETH);
+}
+
+/* This function expects packet with ethernet header with correct
+ * l3 pointer set. */
+static void *
+compose_ipv6(struct dp_packet *packet, uint8_t proto,
+ const struct in6_addr *src, const struct in6_addr *dst,
+ uint8_t key_tc, ovs_be32 key_fl, uint8_t key_hl, int size)
+{
+ struct ip6_hdr *nh;
+ void *data;
+
+ nh = dp_packet_l3(packet);
+ nh->ip6_vfc = 0x60;
+ nh->ip6_nxt = proto;
+ nh->ip6_plen = htons(size);
+ data = dp_packet_put_zeros(packet, size);
+ dp_packet_set_l4(packet, data);
+ packet_set_ipv6(packet, src, dst, key_tc, key_fl, key_hl);
+ return data;
+}
+
+/* Compose an IPv6 Neighbor Discovery Neighbor Solicitation message. */
+void
+compose_nd_ns(struct dp_packet *b, const struct eth_addr eth_src,
+ const struct in6_addr *ipv6_src, const struct in6_addr *ipv6_dst)
+{
+ struct in6_addr sn_addr;
+ struct eth_addr eth_dst;
+ struct ovs_nd_msg *ns;
+ struct ovs_nd_lla_opt *lla_opt;
+ uint32_t icmp_csum;
+
+ in6_addr_solicited_node(&sn_addr, ipv6_dst);
+ ipv6_multicast_to_ethernet(ð_dst, &sn_addr);
+
+ eth_compose(b, eth_dst, eth_src, ETH_TYPE_IPV6, IPV6_HEADER_LEN);
+ ns = compose_ipv6(b, IPPROTO_ICMPV6, ipv6_src, &sn_addr,
+ 0, 0, 255, ND_MSG_LEN + ND_LLA_OPT_LEN);
+
+ ns->icmph.icmp6_type = ND_NEIGHBOR_SOLICIT;
+ ns->icmph.icmp6_code = 0;
+ put_16aligned_be32(&ns->rso_flags, htonl(0));
+
+ lla_opt = &ns->options[0];
+ lla_opt->type = ND_OPT_SOURCE_LINKADDR;
+ lla_opt->len = 1;
+
+ packet_set_nd(b, ipv6_dst, eth_src, eth_addr_zero);
+
+ ns->icmph.icmp6_cksum = 0;
+ icmp_csum = packet_csum_pseudoheader6(dp_packet_l3(b));
+ ns->icmph.icmp6_cksum = csum_finish(
+ csum_continue(icmp_csum, ns, ND_MSG_LEN + ND_LLA_OPT_LEN));
+}
+
+/* Compose an IPv6 Neighbor Discovery Neighbor Advertisement message. */
+void
+compose_nd_na(struct dp_packet *b,
+ const struct eth_addr eth_src, const struct eth_addr eth_dst,
+ const struct in6_addr *ipv6_src, const struct in6_addr *ipv6_dst,
+ ovs_be32 rso_flags)
+{
+ struct ovs_nd_msg *na;
+ struct ovs_nd_lla_opt *lla_opt;
+ uint32_t icmp_csum;
+
+ eth_compose(b, eth_dst, eth_src, ETH_TYPE_IPV6, IPV6_HEADER_LEN);
+ na = compose_ipv6(b, IPPROTO_ICMPV6, ipv6_src, ipv6_dst,
+ 0, 0, 255, ND_MSG_LEN + ND_LLA_OPT_LEN);
+
+ na->icmph.icmp6_type = ND_NEIGHBOR_ADVERT;
+ na->icmph.icmp6_code = 0;
+ put_16aligned_be32(&na->rso_flags, rso_flags);
+
+ lla_opt = &na->options[0];
+ lla_opt->type = ND_OPT_TARGET_LINKADDR;
+ lla_opt->len = 1;
+
+ packet_set_nd(b, ipv6_src, eth_addr_zero, eth_src);
+
+ na->icmph.icmp6_cksum = 0;
+ icmp_csum = packet_csum_pseudoheader6(dp_packet_l3(b));
+ na->icmph.icmp6_cksum = csum_finish(csum_continue(
+ icmp_csum, na, ND_MSG_LEN + ND_LLA_OPT_LEN));
+}
+
+/* Compose an IPv6 Neighbor Discovery Router Advertisement message with
+ * Source Link-layer Address Option and MTU Option.
+ * Caller can call packet_put_ra_prefix_opt to append Prefix Information
+ * Options to composed messags in 'b'. */
+void
+compose_nd_ra(struct dp_packet *b,
+ const struct eth_addr eth_src, const struct eth_addr eth_dst,
+ const struct in6_addr *ipv6_src, const struct in6_addr *ipv6_dst,
+ uint8_t cur_hop_limit, uint8_t mo_flags,
+ ovs_be16 router_lt, ovs_be32 reachable_time,
+ ovs_be32 retrans_timer, ovs_be32 mtu)
+{
+ /* Don't compose Router Advertisement packet with MTU Option if mtu
+ * value is 0. */
+ bool with_mtu = mtu != 0;
+ size_t mtu_opt_len = with_mtu ? ND_MTU_OPT_LEN : 0;
+
+ eth_compose(b, eth_dst, eth_src, ETH_TYPE_IPV6, IPV6_HEADER_LEN);
+
+ struct ovs_ra_msg *ra = compose_ipv6(
+ b, IPPROTO_ICMPV6, ipv6_src, ipv6_dst, 0, 0, 255,
+ RA_MSG_LEN + ND_LLA_OPT_LEN + mtu_opt_len);
+ ra->icmph.icmp6_type = ND_ROUTER_ADVERT;
+ ra->icmph.icmp6_code = 0;
+ ra->cur_hop_limit = cur_hop_limit;
+ ra->mo_flags = mo_flags;
+ ra->router_lifetime = router_lt;
+ ra->reachable_time = reachable_time;
+ ra->retrans_timer = retrans_timer;
+
+ struct ovs_nd_lla_opt *lla_opt = ra->options;
+ lla_opt->type = ND_OPT_SOURCE_LINKADDR;
+ lla_opt->len = 1;
+ lla_opt->mac = eth_src;
+
+ if (with_mtu) {
+ /* ovs_nd_mtu_opt has the same size with ovs_nd_lla_opt. */
+ struct ovs_nd_mtu_opt *mtu_opt
+ = (struct ovs_nd_mtu_opt *)(lla_opt + 1);
+ mtu_opt->type = ND_OPT_MTU;
+ mtu_opt->len = 1;
+ mtu_opt->reserved = 0;
+ put_16aligned_be32(&mtu_opt->mtu, mtu);
+ }
+
+ ra->icmph.icmp6_cksum = 0;
+ uint32_t icmp_csum = packet_csum_pseudoheader6(dp_packet_l3(b));
+ ra->icmph.icmp6_cksum = csum_finish(csum_continue(
+ icmp_csum, ra, RA_MSG_LEN + ND_LLA_OPT_LEN + mtu_opt_len));
+}
+
+/* Append an IPv6 Neighbor Discovery Prefix Information option to a
+ * Router Advertisement message. */
+void
+packet_put_ra_prefix_opt(struct dp_packet *b,
+ uint8_t plen, uint8_t la_flags,
+ ovs_be32 valid_lifetime, ovs_be32 preferred_lifetime,
+ const ovs_be128 prefix)
+{
+ size_t prev_l4_size = dp_packet_l4_size(b);
+ struct ip6_hdr *nh = dp_packet_l3(b);
+ nh->ip6_plen = htons(prev_l4_size + ND_PREFIX_OPT_LEN);
+
+ struct ovs_ra_msg *ra = dp_packet_l4(b);
+ struct ovs_nd_prefix_opt *prefix_opt =
+ dp_packet_put_uninit(b, sizeof *prefix_opt);
+ prefix_opt->type = ND_OPT_PREFIX_INFORMATION;
+ prefix_opt->len = 4;
+ prefix_opt->prefix_len = plen;
+ prefix_opt->la_flags = la_flags;
+ put_16aligned_be32(&prefix_opt->valid_lifetime, valid_lifetime);
+ put_16aligned_be32(&prefix_opt->preferred_lifetime, preferred_lifetime);
+ put_16aligned_be32(&prefix_opt->reserved, 0);
+ memcpy(prefix_opt->prefix.be32, prefix.be32, sizeof(ovs_be32[4]));
+
+ ra->icmph.icmp6_cksum = 0;
+ uint32_t icmp_csum = packet_csum_pseudoheader6(dp_packet_l3(b));
+ ra->icmph.icmp6_cksum = csum_finish(csum_continue(
+ icmp_csum, ra, prev_l4_size + ND_PREFIX_OPT_LEN));
}
uint32_t
return partial;
}
+#ifndef __CHECKER__
+uint32_t
+packet_csum_pseudoheader6(const struct ovs_16aligned_ip6_hdr *ip6)
+{
+ uint32_t partial = 0;
+
+ partial = csum_continue(partial, &ip6->ip6_src, sizeof ip6->ip6_src);
+ partial = csum_continue(partial, &ip6->ip6_dst, sizeof ip6->ip6_dst);
+ partial = csum_add16(partial, htons(ip6->ip6_nxt));
+ partial = csum_add16(partial, ip6->ip6_plen);
+
+ return partial;
+}
+
+/* Calculate the IPv6 upper layer checksum according to RFC2460. We pass the
+ ip6_nxt and ip6_plen values, so it will also work if extension headers
+ are present. */
+uint16_t
+packet_csum_upperlayer6(const struct ovs_16aligned_ip6_hdr *ip6,
+ const void *data, uint8_t l4_protocol,
+ uint16_t l4_size)
+{
+ uint32_t partial = 0;
+
+ partial = csum_continue(partial, &ip6->ip6_src, sizeof ip6->ip6_src);
+ partial = csum_continue(partial, &ip6->ip6_dst, sizeof ip6->ip6_dst);
+ partial = csum_add16(partial, htons(l4_protocol));
+ partial = csum_add16(partial, htons(l4_size));
+
+ partial = csum_continue(partial, data, l4_size);
+
+ return csum_finish(partial);
+}
+#endif
+
+void
+IP_ECN_set_ce(struct dp_packet *pkt, bool is_ipv6)
+{
+ if (is_ipv6) {
+ ovs_16aligned_be32 *ip6 = dp_packet_l3(pkt);
+
+ put_16aligned_be32(ip6, get_16aligned_be32(ip6) |
+ htonl(IP_ECN_CE << 20));
+ } else {
+ struct ip_header *nh = dp_packet_l3(pkt);
+ uint8_t tos = nh->ip_tos;
+
+ tos |= IP_ECN_CE;
+ if (nh->ip_tos != tos) {
+ nh->ip_csum = recalc_csum16(nh->ip_csum, htons(nh->ip_tos),
+ htons((uint16_t) tos));
+ nh->ip_tos = tos;
+ }
+ }
+}