X-Git-Url: http://git.cascardo.eti.br/?a=blobdiff_plain;f=lib%2Fpackets.h;h=bf12937b1005b02648028b5eda4ab04db40ad32c;hb=HEAD;hp=b146a5069e25ed92b8274413278a1c9b770fb704;hpb=8fc55661c57b26ab84a2baa5ef235f1a79c20633;p=cascardo%2Fovs.git diff --git a/lib/packets.h b/lib/packets.h index b146a5069..bf12937b1 100644 --- a/lib/packets.h +++ b/lib/packets.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2008, 2009, 2010, 2011, 2012, 2013, 2014 Nicira, Inc. + * Copyright (c) 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015 Nicira, Inc. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. @@ -23,9 +23,13 @@ #include #include #include "compiler.h" +#include "geneve.h" #include "openvswitch/types.h" +#include "odp-netlink.h" #include "random.h" #include "hash.h" +#include "tun-metadata.h" +#include "unaligned.h" #include "util.h" struct dp_packet; @@ -33,9 +37,11 @@ struct ds; /* Tunnel information used in flow key and metadata. */ struct flow_tnl { - ovs_be64 tun_id; - ovs_be32 ip_src; ovs_be32 ip_dst; + struct in6_addr ipv6_dst; + ovs_be32 ip_src; + struct in6_addr ipv6_src; + ovs_be64 tun_id; uint16_t flags; uint8_t ip_tos; uint8_t ip_ttl; @@ -44,8 +50,79 @@ struct flow_tnl { ovs_be16 gbp_id; uint8_t gbp_flags; uint8_t pad1[5]; /* Pad to 64 bits. */ + struct tun_metadata metadata; }; +/* Some flags are exposed through OpenFlow while others are used only + * internally. */ + +/* Public flags */ +#define FLOW_TNL_F_OAM (1 << 0) + +#define FLOW_TNL_PUB_F_MASK ((1 << 1) - 1) + +/* Private flags */ +#define FLOW_TNL_F_DONT_FRAGMENT (1 << 1) +#define FLOW_TNL_F_CSUM (1 << 2) +#define FLOW_TNL_F_KEY (1 << 3) + +#define FLOW_TNL_F_MASK ((1 << 4) - 1) + +/* Purely internal to OVS userspace. These flags should never be exposed to + * the outside world and so aren't included in the flags mask. */ + +/* Tunnel information is in userspace datapath format. */ +#define FLOW_TNL_F_UDPIF (1 << 4) + +static inline bool ipv6_addr_is_set(const struct in6_addr *addr); + +static inline bool +flow_tnl_dst_is_set(const struct flow_tnl *tnl) +{ + return tnl->ip_dst || ipv6_addr_is_set(&tnl->ipv6_dst); +} + +struct in6_addr flow_tnl_dst(const struct flow_tnl *tnl); +struct in6_addr flow_tnl_src(const struct flow_tnl *tnl); + +/* Returns an offset to 'src' covering all the meaningful fields in 'src'. */ +static inline size_t +flow_tnl_size(const struct flow_tnl *src) +{ + if (!flow_tnl_dst_is_set(src)) { + /* Covers ip_dst and ipv6_dst only. */ + return offsetof(struct flow_tnl, ip_src); + } + if (src->flags & FLOW_TNL_F_UDPIF) { + /* Datapath format, cover all options we have. */ + return offsetof(struct flow_tnl, metadata.opts) + + src->metadata.present.len; + } + if (!src->metadata.present.map) { + /* No TLVs, opts is irrelevant. */ + return offsetof(struct flow_tnl, metadata.opts); + } + /* Have decoded TLVs, opts is relevant. */ + return sizeof *src; +} + +/* Copy flow_tnl, but avoid copying unused portions of tun_metadata. Unused + * data in 'dst' is NOT cleared, so this must not be used in cases where the + * uninitialized portion may be hashed over. */ +static inline void +flow_tnl_copy__(struct flow_tnl *dst, const struct flow_tnl *src) +{ + memcpy(dst, src, flow_tnl_size(src)); +} + +static inline bool +flow_tnl_equal(const struct flow_tnl *a, const struct flow_tnl *b) +{ + size_t a_size = flow_tnl_size(a); + + return a_size == flow_tnl_size(b) && !memcmp(a, b, a_size); +} + /* Unfortunately, a "struct flow" sometimes has to handle OpenFlow port * numbers and other times datapath (dpif) port numbers. This union allows * access to both. */ @@ -61,143 +138,182 @@ struct pkt_metadata { received from the wire. */ uint32_t dp_hash; /* hash value computed by the recirculation action. */ - struct flow_tnl tunnel; /* Encapsulating tunnel parameters. */ uint32_t skb_priority; /* Packet priority for QoS. */ uint32_t pkt_mark; /* Packet mark. */ + uint16_t ct_state; /* Connection state. */ + uint16_t ct_zone; /* Connection zone. */ + uint32_t ct_mark; /* Connection mark. */ + ovs_u128 ct_label; /* Connection label. */ union flow_in_port in_port; /* Input port. */ + struct flow_tnl tunnel; /* Encapsulating tunnel parameters. Note that + * if 'ip_dst' == 0, the rest of the fields may + * be uninitialized. */ }; -#define PKT_METADATA_INITIALIZER(PORT) \ - (struct pkt_metadata){ .in_port.odp_port = PORT } +static inline void +pkt_metadata_init(struct pkt_metadata *md, odp_port_t port) +{ + /* It can be expensive to zero out all of the tunnel metadata. However, + * we can just zero out ip_dst and the rest of the data will never be + * looked at. */ + memset(md, 0, offsetof(struct pkt_metadata, in_port)); + md->tunnel.ip_dst = 0; + md->tunnel.ipv6_dst = in6addr_any; + + md->in_port.odp_port = port; +} + +/* This function prefetches the cachelines touched by pkt_metadata_init() + * For performance reasons the two functions should be kept in sync. */ +static inline void +pkt_metadata_prefetch_init(struct pkt_metadata *md) +{ + ovs_prefetch_range(md, offsetof(struct pkt_metadata, tunnel.ip_src)); +} bool dpid_from_string(const char *s, uint64_t *dpidp); #define ETH_ADDR_LEN 6 -static const uint8_t eth_addr_broadcast[ETH_ADDR_LEN] OVS_UNUSED - = { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff }; +static const struct eth_addr eth_addr_broadcast OVS_UNUSED + = { { { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff } } }; + +static const struct eth_addr eth_addr_exact OVS_UNUSED + = { { { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff } } }; + +static const struct eth_addr eth_addr_zero OVS_UNUSED + = { { { 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 } } }; -static const uint8_t eth_addr_stp[ETH_ADDR_LEN] OVS_UNUSED - = { 0x01, 0x80, 0xC2, 0x00, 0x00, 0x00 }; +static const struct eth_addr eth_addr_stp OVS_UNUSED + = { { { 0x01, 0x80, 0xC2, 0x00, 0x00, 0x00 } } }; -static const uint8_t eth_addr_lacp[ETH_ADDR_LEN] OVS_UNUSED - = { 0x01, 0x80, 0xC2, 0x00, 0x00, 0x02 }; +static const struct eth_addr eth_addr_lacp OVS_UNUSED + = { { { 0x01, 0x80, 0xC2, 0x00, 0x00, 0x02 } } }; -static const uint8_t eth_addr_bfd[ETH_ADDR_LEN] OVS_UNUSED - = { 0x00, 0x23, 0x20, 0x00, 0x00, 0x01 }; +static const struct eth_addr eth_addr_bfd OVS_UNUSED + = { { { 0x00, 0x23, 0x20, 0x00, 0x00, 0x01 } } }; -static inline bool eth_addr_is_broadcast(const uint8_t ea[ETH_ADDR_LEN]) +static inline bool eth_addr_is_broadcast(const struct eth_addr a) { - return (ea[0] & ea[1] & ea[2] & ea[3] & ea[4] & ea[5]) == 0xff; + return (a.be16[0] & a.be16[1] & a.be16[2]) == htons(0xffff); } -static inline bool eth_addr_is_multicast(const uint8_t ea[ETH_ADDR_LEN]) +static inline bool eth_addr_is_multicast(const struct eth_addr a) { - return ea[0] & 1; + return a.ea[0] & 1; } -static inline bool eth_addr_is_local(const uint8_t ea[ETH_ADDR_LEN]) + +static inline bool eth_addr_is_local(const struct eth_addr a) { /* Local if it is either a locally administered address or a Nicira random * address. */ - return ea[0] & 2 - || (ea[0] == 0x00 && ea[1] == 0x23 && ea[2] == 0x20 && ea[3] & 0x80); + return a.ea[0] & 2 + || (a.be16[0] == htons(0x0023) + && (a.be16[1] & htons(0xff80)) == htons(0x2080)); } -static inline bool eth_addr_is_zero(const uint8_t ea[ETH_ADDR_LEN]) +static inline bool eth_addr_is_zero(const struct eth_addr a) { - return !(ea[0] | ea[1] | ea[2] | ea[3] | ea[4] | ea[5]); + return !(a.be16[0] | a.be16[1] | a.be16[2]); } -static inline int eth_mask_is_exact(const uint8_t ea[ETH_ADDR_LEN]) +static inline int eth_mask_is_exact(const struct eth_addr a) { - return (ea[0] & ea[1] & ea[2] & ea[3] & ea[4] & ea[5]) == 0xff; + return (a.be16[0] & a.be16[1] & a.be16[2]) == htons(0xffff); } -static inline int eth_addr_compare_3way(const uint8_t a[ETH_ADDR_LEN], - const uint8_t b[ETH_ADDR_LEN]) +static inline int eth_addr_compare_3way(const struct eth_addr a, + const struct eth_addr b) { - return memcmp(a, b, ETH_ADDR_LEN); + return memcmp(&a, &b, sizeof a); } -static inline bool eth_addr_equals(const uint8_t a[ETH_ADDR_LEN], - const uint8_t b[ETH_ADDR_LEN]) + +static inline bool eth_addr_equals(const struct eth_addr a, + const struct eth_addr b) { return !eth_addr_compare_3way(a, b); } -static inline bool eth_addr_equal_except(const uint8_t a[ETH_ADDR_LEN], - const uint8_t b[ETH_ADDR_LEN], - const uint8_t mask[ETH_ADDR_LEN]) + +static inline bool eth_addr_equal_except(const struct eth_addr a, + const struct eth_addr b, + const struct eth_addr mask) { - return !(((a[0] ^ b[0]) & mask[0]) - || ((a[1] ^ b[1]) & mask[1]) - || ((a[2] ^ b[2]) & mask[2]) - || ((a[3] ^ b[3]) & mask[3]) - || ((a[4] ^ b[4]) & mask[4]) - || ((a[5] ^ b[5]) & mask[5])); + return !(((a.be16[0] ^ b.be16[0]) & mask.be16[0]) + || ((a.be16[1] ^ b.be16[1]) & mask.be16[1]) + || ((a.be16[2] ^ b.be16[2]) & mask.be16[2])); } -static inline uint64_t eth_addr_to_uint64(const uint8_t ea[ETH_ADDR_LEN]) + +static inline uint64_t eth_addr_to_uint64(const struct eth_addr ea) { - return (((uint64_t) ea[0] << 40) - | ((uint64_t) ea[1] << 32) - | ((uint64_t) ea[2] << 24) - | ((uint64_t) ea[3] << 16) - | ((uint64_t) ea[4] << 8) - | ea[5]); + return (((uint64_t) ntohs(ea.be16[0]) << 32) + | ((uint64_t) ntohs(ea.be16[1]) << 16) + | ntohs(ea.be16[2])); } -static inline uint64_t eth_addr_vlan_to_uint64(const uint8_t ea[ETH_ADDR_LEN], + +static inline uint64_t eth_addr_vlan_to_uint64(const struct eth_addr ea, uint16_t vlan) { return (((uint64_t)vlan << 48) | eth_addr_to_uint64(ea)); } -static inline void eth_addr_from_uint64(uint64_t x, uint8_t ea[ETH_ADDR_LEN]) + +static inline void eth_addr_from_uint64(uint64_t x, struct eth_addr *ea) { - ea[0] = x >> 40; - ea[1] = x >> 32; - ea[2] = x >> 24; - ea[3] = x >> 16; - ea[4] = x >> 8; - ea[5] = x; + ea->be16[0] = htons(x >> 32); + ea->be16[1] = htons((x & 0xFFFF0000) >> 16); + ea->be16[2] = htons(x & 0xFFFF); } -static inline void eth_addr_mark_random(uint8_t ea[ETH_ADDR_LEN]) + +static inline struct eth_addr eth_addr_invert(const struct eth_addr src) { - ea[0] &= ~1; /* Unicast. */ - ea[0] |= 2; /* Private. */ + struct eth_addr dst; + + for (int i = 0; i < ARRAY_SIZE(src.be16); i++) { + dst.be16[i] = ~src.be16[i]; + } + + return dst; } -static inline void eth_addr_random(uint8_t ea[ETH_ADDR_LEN]) + +static inline void eth_addr_mark_random(struct eth_addr *ea) { - random_bytes(ea, ETH_ADDR_LEN); + ea->ea[0] &= ~1; /* Unicast. */ + ea->ea[0] |= 2; /* Private. */ +} + +static inline void eth_addr_random(struct eth_addr *ea) +{ + random_bytes((uint8_t *)ea, sizeof *ea); eth_addr_mark_random(ea); } -static inline void eth_addr_nicira_random(uint8_t ea[ETH_ADDR_LEN]) + +static inline void eth_addr_nicira_random(struct eth_addr *ea) { eth_addr_random(ea); /* Set the OUI to the Nicira one. */ - ea[0] = 0x00; - ea[1] = 0x23; - ea[2] = 0x20; + ea->ea[0] = 0x00; + ea->ea[1] = 0x23; + ea->ea[2] = 0x20; /* Set the top bit to indicate random Nicira address. */ - ea[3] |= 0x80; + ea->ea[3] |= 0x80; } -static inline uint32_t hash_mac(const uint8_t ea[ETH_ADDR_LEN], +static inline uint32_t hash_mac(const struct eth_addr ea, const uint16_t vlan, const uint32_t basis) { return hash_uint64_basis(eth_addr_vlan_to_uint64(ea, vlan), basis); } -bool eth_addr_is_reserved(const uint8_t ea[ETH_ADDR_LEN]); -bool eth_addr_from_string(const char *, uint8_t ea[ETH_ADDR_LEN]); +bool eth_addr_is_reserved(const struct eth_addr); +bool eth_addr_from_string(const char *, struct eth_addr *); -void compose_rarp(struct dp_packet *, const uint8_t eth_src[ETH_ADDR_LEN]); +void compose_rarp(struct dp_packet *, const struct eth_addr); void eth_push_vlan(struct dp_packet *, ovs_be16 tpid, ovs_be16 tci); void eth_pop_vlan(struct dp_packet *); const char *eth_from_hex(const char *hex, struct dp_packet **packetp); -void eth_format_masked(const uint8_t eth[ETH_ADDR_LEN], - const uint8_t mask[ETH_ADDR_LEN], struct ds *s); -void eth_addr_bitand(const uint8_t src[ETH_ADDR_LEN], - const uint8_t mask[ETH_ADDR_LEN], - uint8_t dst[ETH_ADDR_LEN]); +void eth_format_masked(const struct eth_addr ea, + const struct eth_addr *mask, struct ds *s); void set_mpls_lse(struct dp_packet *, ovs_be32 label); void push_mpls(struct dp_packet *packet, ovs_be16 ethtype, ovs_be32 lse); @@ -212,20 +328,22 @@ ovs_be32 set_mpls_lse_values(uint8_t ttl, uint8_t tc, uint8_t bos, /* Example: * - * uint8_t mac[ETH_ADDR_LEN]; + * struct eth_addr mac; * [...] * printf("The Ethernet address is "ETH_ADDR_FMT"\n", ETH_ADDR_ARGS(mac)); * */ #define ETH_ADDR_FMT \ "%02"PRIx8":%02"PRIx8":%02"PRIx8":%02"PRIx8":%02"PRIx8":%02"PRIx8 -#define ETH_ADDR_ARGS(ea) \ - (ea)[0], (ea)[1], (ea)[2], (ea)[3], (ea)[4], (ea)[5] +#define ETH_ADDR_ARGS(EA) ETH_ADDR_BYTES_ARGS((EA).ea) +#define ETH_ADDR_BYTES_ARGS(EAB) \ + (EAB)[0], (EAB)[1], (EAB)[2], (EAB)[3], (EAB)[4], (EAB)[5] +#define ETH_ADDR_STRLEN 17 /* Example: * * char *string = "1 00:11:22:33:44:55 2"; - * uint8_t mac[ETH_ADDR_LEN]; + * struct eth_addr mac; * int a, b; * * if (ovs_scan(string, "%d"ETH_ADDR_SCAN_FMT"%d", @@ -234,8 +352,8 @@ ovs_be32 set_mpls_lse_values(uint8_t ttl, uint8_t tc, uint8_t bos, * } */ #define ETH_ADDR_SCAN_FMT "%"SCNx8":%"SCNx8":%"SCNx8":%"SCNx8":%"SCNx8":%"SCNx8 -#define ETH_ADDR_SCAN_ARGS(ea) \ - &(ea)[0], &(ea)[1], &(ea)[2], &(ea)[3], &(ea)[4], &(ea)[5] +#define ETH_ADDR_SCAN_ARGS(EA) \ + &(EA).ea[0], &(EA).ea[1], &(EA).ea[2], &(EA).ea[3], &(EA).ea[4], &(EA).ea[5] #define ETH_TYPE_IP 0x0800 #define ETH_TYPE_ARP 0x0806 @@ -255,6 +373,13 @@ static inline bool eth_type_mpls(ovs_be16 eth_type) eth_type == htons(ETH_TYPE_MPLS_MCAST); } +static inline bool eth_type_vlan(ovs_be16 eth_type) +{ + return eth_type == htons(ETH_TYPE_VLAN_8021Q) || + eth_type == htons(ETH_TYPE_VLAN_8021AD); +} + + /* Minimum value for an Ethernet type. Values below this are IEEE 802.2 frame * lengths. */ #define ETH_TYPE_MIN 0x600 @@ -267,8 +392,8 @@ static inline bool eth_type_mpls(ovs_be16 eth_type) #define ETH_VLAN_TOTAL_MAX (ETH_HEADER_LEN + VLAN_HEADER_LEN + ETH_PAYLOAD_MAX) OVS_PACKED( struct eth_header { - uint8_t eth_dst[ETH_ADDR_LEN]; - uint8_t eth_src[ETH_ADDR_LEN]; + struct eth_addr eth_dst; + struct eth_addr eth_src; ovs_be16 eth_type; }); BUILD_ASSERT_DECL(ETH_HEADER_LEN == sizeof(struct eth_header)); @@ -352,8 +477,8 @@ BUILD_ASSERT_DECL(VLAN_HEADER_LEN == sizeof(struct vlan_header)); #define VLAN_ETH_HEADER_LEN (ETH_HEADER_LEN + VLAN_HEADER_LEN) OVS_PACKED( struct vlan_eth_header { - uint8_t veth_dst[ETH_ADDR_LEN]; - uint8_t veth_src[ETH_ADDR_LEN]; + struct eth_addr veth_dst; + struct eth_addr veth_src; ovs_be16 veth_type; /* Always htons(ETH_TYPE_VLAN). */ ovs_be16 veth_tci; /* Lowest 12 bits are VLAN ID. */ ovs_be16 veth_next_type; @@ -464,6 +589,16 @@ ip_is_local_multicast(ovs_be32 ip) } int ip_count_cidr_bits(ovs_be32 netmask); void ip_format_masked(ovs_be32 ip, ovs_be32 mask, struct ds *); +bool ip_parse(const char *s, ovs_be32 *ip); +char *ip_parse_masked(const char *s, ovs_be32 *ip, ovs_be32 *mask) + OVS_WARN_UNUSED_RESULT; +char *ip_parse_cidr(const char *s, ovs_be32 *ip, unsigned int *plen) + OVS_WARN_UNUSED_RESULT; +char *ip_parse_masked_len(const char *s, int *n, ovs_be32 *ip, ovs_be32 *mask) + OVS_WARN_UNUSED_RESULT; +char *ip_parse_cidr_len(const char *s, int *n, ovs_be32 *ip, + unsigned int *plen) + OVS_WARN_UNUSED_RESULT; #define IP_VER(ip_ihl_ver) ((ip_ihl_ver) >> 4) #define IP_IHL(ip_ihl_ver) ((ip_ihl_ver) & 15) @@ -533,12 +668,41 @@ struct igmp_header { }; BUILD_ASSERT_DECL(IGMP_HEADER_LEN == sizeof(struct igmp_header)); +#define IGMPV3_HEADER_LEN 8 +struct igmpv3_header { + uint8_t type; + uint8_t rsvr1; + ovs_be16 csum; + ovs_be16 rsvr2; + ovs_be16 ngrp; +}; +BUILD_ASSERT_DECL(IGMPV3_HEADER_LEN == sizeof(struct igmpv3_header)); + +#define IGMPV3_RECORD_LEN 8 +struct igmpv3_record { + uint8_t type; + uint8_t aux_len; + ovs_be16 nsrcs; + ovs_16aligned_be32 maddr; +}; +BUILD_ASSERT_DECL(IGMPV3_RECORD_LEN == sizeof(struct igmpv3_record)); + #define IGMP_HOST_MEMBERSHIP_QUERY 0x11 /* From RFC1112 */ #define IGMP_HOST_MEMBERSHIP_REPORT 0x12 /* Ditto */ #define IGMPV2_HOST_MEMBERSHIP_REPORT 0x16 /* V2 version of 0x12 */ #define IGMP_HOST_LEAVE_MESSAGE 0x17 #define IGMPV3_HOST_MEMBERSHIP_REPORT 0x22 /* V3 version of 0x12 */ +/* + * IGMPv3 and MLDv2 use the same codes. + */ +#define IGMPV3_MODE_IS_INCLUDE 1 +#define IGMPV3_MODE_IS_EXCLUDE 2 +#define IGMPV3_CHANGE_TO_INCLUDE_MODE 3 +#define IGMPV3_CHANGE_TO_EXCLUDE_MODE 4 +#define IGMPV3_ALLOW_NEW_SOURCES 5 +#define IGMPV3_BLOCK_OLD_SOURCES 6 + #define SCTP_HEADER_LEN 12 struct sctp_header { ovs_be16 sctp_src; @@ -585,6 +749,35 @@ struct tcp_header { }; BUILD_ASSERT_DECL(TCP_HEADER_LEN == sizeof(struct tcp_header)); +/* Connection states */ +enum { + CS_NEW_BIT = 0, + CS_ESTABLISHED_BIT = 1, + CS_RELATED_BIT = 2, + CS_REPLY_DIR_BIT = 3, + CS_INVALID_BIT = 4, + CS_TRACKED_BIT = 5, + CS_SRC_NAT_BIT = 6, + CS_DST_NAT_BIT = 7, +}; + +enum { + CS_NEW = (1 << CS_NEW_BIT), + CS_ESTABLISHED = (1 << CS_ESTABLISHED_BIT), + CS_RELATED = (1 << CS_RELATED_BIT), + CS_REPLY_DIR = (1 << CS_REPLY_DIR_BIT), + CS_INVALID = (1 << CS_INVALID_BIT), + CS_TRACKED = (1 << CS_TRACKED_BIT), + CS_SRC_NAT = (1 << CS_SRC_NAT_BIT), + CS_DST_NAT = (1 << CS_DST_NAT_BIT), +}; + +/* Undefined connection state bits. */ +#define CS_SUPPORTED_MASK (CS_NEW | CS_ESTABLISHED | CS_RELATED \ + | CS_INVALID | CS_REPLY_DIR | CS_TRACKED \ + | CS_SRC_NAT | CS_DST_NAT) +#define CS_UNSUPPORTED_MASK (~(uint32_t)CS_SUPPORTED_MASK) + #define ARP_HRD_ETHERNET 1 #define ARP_PRO_IP 0x0800 #define ARP_OP_REQUEST 1 @@ -601,13 +794,15 @@ struct arp_eth_header { ovs_be16 ar_op; /* Opcode. */ /* Ethernet+IPv4 specific members. */ - uint8_t ar_sha[ETH_ADDR_LEN]; /* Sender hardware address. */ - ovs_16aligned_be32 ar_spa; /* Sender protocol address. */ - uint8_t ar_tha[ETH_ADDR_LEN]; /* Target hardware address. */ - ovs_16aligned_be32 ar_tpa; /* Target protocol address. */ + struct eth_addr ar_sha; /* Sender hardware address. */ + ovs_16aligned_be32 ar_spa; /* Sender protocol address. */ + struct eth_addr ar_tha; /* Target hardware address. */ + ovs_16aligned_be32 ar_tpa; /* Target protocol address. */ }; BUILD_ASSERT_DECL(ARP_ETH_HEADER_LEN == sizeof(struct arp_eth_header)); +#define IPV6_HEADER_LEN 40 + /* Like struct in6_addr, but whereas that struct requires 32-bit alignment on * most implementations, this one only requires 16-bit alignment. */ union ovs_16aligned_in6_addr { @@ -648,13 +843,15 @@ struct icmp6_header { }; BUILD_ASSERT_DECL(ICMP6_HEADER_LEN == sizeof(struct icmp6_header)); +uint32_t packet_csum_pseudoheader6(const struct ovs_16aligned_ip6_hdr *); + /* Neighbor Discovery option field. * ND options are always a multiple of 8 bytes in size. */ #define ND_OPT_LEN 8 struct ovs_nd_opt { uint8_t nd_opt_type; /* Values defined in icmp6.h */ uint8_t nd_opt_len; /* in units of 8 octets (the size of this struct) */ - uint8_t nd_opt_data[6]; /* Ethernet address in the case of SLL or TLL options */ + struct eth_addr nd_opt_mac; /* Ethernet address in the case of SLL or TLL options */ }; BUILD_ASSERT_DECL(ND_OPT_LEN == sizeof(struct ovs_nd_opt)); @@ -669,6 +866,35 @@ struct ovs_nd_msg { }; BUILD_ASSERT_DECL(ND_MSG_LEN == sizeof(struct ovs_nd_msg)); +/* + * Use the same struct for MLD and MLD2, naming members as the defined fields in + * in the corresponding version of the protocol, though they are reserved in the + * other one. + */ +#define MLD_HEADER_LEN 8 +struct mld_header { + uint8_t type; + uint8_t code; + ovs_be16 csum; + ovs_be16 mrd; + ovs_be16 ngrp; +}; +BUILD_ASSERT_DECL(MLD_HEADER_LEN == sizeof(struct mld_header)); + +#define MLD2_RECORD_LEN 20 +struct mld2_record { + uint8_t type; + uint8_t aux_len; + ovs_be16 nsrcs; + union ovs_16aligned_in6_addr maddr; +}; +BUILD_ASSERT_DECL(MLD2_RECORD_LEN == sizeof(struct mld2_record)); + +#define MLD_QUERY 130 +#define MLD_REPORT 131 +#define MLD_DONE 132 +#define MLD2_REPORT 143 + /* The IPv6 flow label is in the lower 20 bits of the first 32-bit word. */ #define IPV6_LABEL_MASK 0x000fffff @@ -690,6 +916,10 @@ extern const struct in6_addr in6addr_exact; #define IN6ADDR_EXACT_INIT { { { 0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff, \ 0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff } } } +extern const struct in6_addr in6addr_all_hosts; +#define IN6ADDR_ALL_HOSTS_INIT { { { 0xff,0x02,0x00,0x00,0x00,0x00,0x00,0x00, \ + 0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x01 } } } + static inline bool ipv6_addr_equals(const struct in6_addr *a, const struct in6_addr *b) { @@ -708,6 +938,65 @@ static inline bool ipv6_mask_is_exact(const struct in6_addr *mask) { return ipv6_addr_equals(mask, &in6addr_exact); } +static inline bool ipv6_is_all_hosts(const struct in6_addr *addr) { + return ipv6_addr_equals(addr, &in6addr_all_hosts); +} + +static inline bool ipv6_addr_is_set(const struct in6_addr *addr) { + return !ipv6_addr_equals(addr, &in6addr_any); +} + +static inline bool ipv6_addr_is_multicast(const struct in6_addr *ip) { + return ip->s6_addr[0] == 0xff; +} + +static inline struct in6_addr +in6_addr_mapped_ipv4(ovs_be32 ip4) +{ + struct in6_addr ip6 = { .s6_addr = { [10] = 0xff, [11] = 0xff } }; + memcpy(&ip6.s6_addr[12], &ip4, 4); + return ip6; +} + +static inline void +in6_addr_set_mapped_ipv4(struct in6_addr *ip6, ovs_be32 ip4) +{ + *ip6 = in6_addr_mapped_ipv4(ip4); +} + +static inline ovs_be32 +in6_addr_get_mapped_ipv4(const struct in6_addr *addr) +{ + union ovs_16aligned_in6_addr *taddr = (void *) addr; + if (IN6_IS_ADDR_V4MAPPED(addr)) { + return get_16aligned_be32(&taddr->be32[3]); + } else { + return INADDR_ANY; + } +} + +static inline void +in6_addr_solicited_node(struct in6_addr *addr, const struct in6_addr *ip6) +{ + union ovs_16aligned_in6_addr *taddr = (void *) addr; + memset(taddr->be16, 0, sizeof(taddr->be16)); + taddr->be16[0] = htons(0xff02); + taddr->be16[5] = htons(0x1); + taddr->be16[6] = htons(0xff00); + memcpy(&addr->s6_addr[13], &ip6->s6_addr[13], 3); +} + +static inline void +ipv6_multicast_to_ethernet(struct eth_addr *eth, const struct in6_addr *ip6) +{ + eth->ea[0] = 0x33; + eth->ea[1] = 0x33; + eth->ea[2] = ip6->s6_addr[12]; + eth->ea[3] = ip6->s6_addr[13]; + eth->ea[4] = ip6->s6_addr[14]; + eth->ea[5] = ip6->s6_addr[15]; +} + static inline bool dl_type_is_ip_any(ovs_be16 dl_type) { return dl_type == htons(ETH_TYPE_IP) @@ -715,42 +1004,6 @@ static inline bool dl_type_is_ip_any(ovs_be16 dl_type) } /* Tunnel header */ -#define GENEVE_CRIT_OPT_TYPE (1 << 7) -struct geneve_opt { - ovs_be16 opt_class; - uint8_t type; -#ifdef WORDS_BIGENDIAN - uint8_t r1:1; - uint8_t r2:1; - uint8_t r3:1; - uint8_t length:5; -#else - uint8_t length:5; - uint8_t r3:1; - uint8_t r2:1; - uint8_t r1:1; -#endif - /* Option data */ -}; - -struct genevehdr { -#ifdef WORDS_BIGENDIAN - uint8_t ver:2; - uint8_t opt_len:6; - uint8_t oam:1; - uint8_t critical:1; - uint8_t rsvd1:6; -#else - uint8_t opt_len:6; - uint8_t ver:2; - uint8_t rsvd1:6; - uint8_t critical:1; - uint8_t oam:1; -#endif - ovs_be16 proto_type; - ovs_16aligned_be32 vni; - struct geneve_opt options[]; -}; /* GRE protocol header */ struct gre_base_hdr { @@ -775,21 +1028,35 @@ struct vxlanhdr { #define VXLAN_FLAGS 0x08000000 /* struct vxlanhdr.vx_flags required value. */ -void format_ipv6_addr(char *addr_str, const struct in6_addr *addr); -void print_ipv6_addr(struct ds *string, const struct in6_addr *addr); -void print_ipv6_masked(struct ds *string, const struct in6_addr *addr, - const struct in6_addr *mask); +void ipv6_format_addr(const struct in6_addr *addr, struct ds *); +void ipv6_format_addr_bracket(const struct in6_addr *addr, struct ds *, + bool bracket); +void ipv6_format_mapped(const struct in6_addr *addr, struct ds *); +void ipv6_format_masked(const struct in6_addr *addr, + const struct in6_addr *mask, struct ds *); +const char * ipv6_string_mapped(char *addr_str, const struct in6_addr *addr); struct in6_addr ipv6_addr_bitand(const struct in6_addr *src, const struct in6_addr *mask); struct in6_addr ipv6_create_mask(int mask); int ipv6_count_cidr_bits(const struct in6_addr *netmask); bool ipv6_is_cidr(const struct in6_addr *netmask); -void *eth_compose(struct dp_packet *, const uint8_t eth_dst[ETH_ADDR_LEN], - const uint8_t eth_src[ETH_ADDR_LEN], uint16_t eth_type, +bool ipv6_parse(const char *s, struct in6_addr *ip); +char *ipv6_parse_masked(const char *s, struct in6_addr *ipv6, + struct in6_addr *mask); +char *ipv6_parse_cidr(const char *s, struct in6_addr *ip, unsigned int *plen) + OVS_WARN_UNUSED_RESULT; +char *ipv6_parse_masked_len(const char *s, int *n, struct in6_addr *ipv6, + struct in6_addr *mask); +char *ipv6_parse_cidr_len(const char *s, int *n, struct in6_addr *ip, + unsigned int *plen) + OVS_WARN_UNUSED_RESULT; + +void *eth_compose(struct dp_packet *, const struct eth_addr eth_dst, + const struct eth_addr eth_src, uint16_t eth_type, size_t size); -void *snap_compose(struct dp_packet *, const uint8_t eth_dst[ETH_ADDR_LEN], - const uint8_t eth_src[ETH_ADDR_LEN], +void *snap_compose(struct dp_packet *, const struct eth_addr eth_dst, + const struct eth_addr eth_src, unsigned int oui, uint16_t snap_type, size_t size); void packet_set_ipv4(struct dp_packet *, ovs_be32 src, ovs_be32 dst, uint8_t tos, uint8_t ttl); @@ -799,13 +1066,18 @@ void packet_set_ipv6(struct dp_packet *, uint8_t proto, const ovs_be32 src[4], void packet_set_tcp_port(struct dp_packet *, ovs_be16 src, ovs_be16 dst); void packet_set_udp_port(struct dp_packet *, ovs_be16 src, ovs_be16 dst); void packet_set_sctp_port(struct dp_packet *, ovs_be16 src, ovs_be16 dst); +void packet_set_icmp(struct dp_packet *, uint8_t type, uint8_t code); void packet_set_nd(struct dp_packet *, const ovs_be32 target[4], - const uint8_t sll[6], const uint8_t tll[6]); + const struct eth_addr sll, const struct eth_addr tll); void packet_format_tcp_flags(struct ds *, uint16_t); const char *packet_tcp_flag_to_string(uint32_t flag); -void compose_arp(struct dp_packet *b, const uint8_t eth_src[ETH_ADDR_LEN], - ovs_be32 ip_src, ovs_be32 ip_dst); +void compose_arp(struct dp_packet *, 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); +void compose_nd(struct dp_packet *, const struct eth_addr eth_src, + struct in6_addr *, struct in6_addr *); uint32_t packet_csum_pseudoheader(const struct ip_header *); #endif /* packets.h */