2
0
mirror of https://github.com/openvswitch/ovs synced 2025-08-22 18:07:40 +00:00
ovs/lib/packets.h

706 lines
21 KiB
C
Raw Normal View History

/*
* Copyright (c) 2008, 2009, 2010, 2011, 2012, 2013, 2014 Nicira, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at:
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#ifndef PACKETS_H
#define PACKETS_H 1
#include <inttypes.h>
#include <sys/types.h>
#include <netinet/in.h>
#include <stdint.h>
#include <string.h>
#include "compiler.h"
#include "openvswitch/types.h"
#include "random.h"
#include "hash.h"
#include "util.h"
struct ofpbuf;
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;
uint16_t flags;
uint8_t ip_tos;
uint8_t ip_ttl;
};
/* Unfortunately, a "struct flow" sometimes has to handle OpenFlow port
* numbers and other times datapath (dpif) port numbers. This union allows
* access to both. */
union flow_in_port {
odp_port_t odp_port;
ofp_port_t ofp_port;
};
/* Datapath packet metadata */
struct pkt_metadata {
uint32_t recirc_id; /* Recirculation id carried with the
recirculating packets. 0 for packets
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. */
union flow_in_port in_port; /* Input port. */
};
#define PKT_METADATA_INITIALIZER(PORT) \
(struct pkt_metadata){ 0, 0, { 0, 0, 0, 0, 0, 0}, 0, 0, {(PORT)} }
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 uint8_t eth_addr_stp[ETH_ADDR_LEN] 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 uint8_t eth_addr_bfd[ETH_ADDR_LEN] OVS_UNUSED
= { 0x00, 0x23, 0x20, 0x00, 0x00, 0x01 };
static inline bool eth_addr_is_broadcast(const uint8_t ea[6])
{
return (ea[0] & ea[1] & ea[2] & ea[3] & ea[4] & ea[5]) == 0xff;
}
static inline bool eth_addr_is_multicast(const uint8_t ea[6])
{
return ea[0] & 1;
}
static inline bool eth_addr_is_local(const uint8_t ea[6])
{
/* 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);
}
static inline bool eth_addr_is_zero(const uint8_t ea[6])
{
return !(ea[0] | ea[1] | ea[2] | ea[3] | ea[4] | ea[5]);
}
static inline int eth_mask_is_exact(const uint8_t ea[ETH_ADDR_LEN])
{
return (ea[0] & ea[1] & ea[2] & ea[3] & ea[4] & ea[5]) == 0xff;
}
static inline int eth_addr_compare_3way(const uint8_t a[ETH_ADDR_LEN],
const uint8_t b[ETH_ADDR_LEN])
{
return memcmp(a, b, ETH_ADDR_LEN);
}
static inline bool eth_addr_equals(const uint8_t a[ETH_ADDR_LEN],
const uint8_t b[ETH_ADDR_LEN])
{
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])
{
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]));
}
static inline uint64_t eth_addr_to_uint64(const uint8_t ea[ETH_ADDR_LEN])
{
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]);
}
static inline uint64_t eth_addr_vlan_to_uint64(const uint8_t ea[ETH_ADDR_LEN],
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])
{
ea[0] = x >> 40;
ea[1] = x >> 32;
ea[2] = x >> 24;
ea[3] = x >> 16;
ea[4] = x >> 8;
ea[5] = x;
}
static inline void eth_addr_mark_random(uint8_t ea[ETH_ADDR_LEN])
{
ea[0] &= ~1; /* Unicast. */
ea[0] |= 2; /* Private. */
}
static inline void eth_addr_random(uint8_t ea[ETH_ADDR_LEN])
{
random_bytes(ea, ETH_ADDR_LEN);
eth_addr_mark_random(ea);
}
static inline void eth_addr_nicira_random(uint8_t ea[ETH_ADDR_LEN])
{
eth_addr_random(ea);
/* Set the OUI to the Nicira one. */
ea[0] = 0x00;
ea[1] = 0x23;
ea[2] = 0x20;
/* Set the top bit to indicate random Nicira address. */
ea[3] |= 0x80;
}
static inline uint32_t hash_mac(const uint8_t ea[ETH_ADDR_LEN],
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]);
void compose_rarp(struct ofpbuf *, const uint8_t eth_src[ETH_ADDR_LEN]);
void eth_push_vlan(struct ofpbuf *, ovs_be16 tpid, ovs_be16 tci);
void eth_pop_vlan(struct ofpbuf *);
const char *eth_from_hex(const char *hex, struct ofpbuf **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 set_mpls_lse(struct ofpbuf *, ovs_be32 label);
void push_mpls(struct ofpbuf *packet, ovs_be16 ethtype, ovs_be32 lse);
void pop_mpls(struct ofpbuf *, ovs_be16 ethtype);
void set_mpls_lse_ttl(ovs_be32 *lse, uint8_t ttl);
void set_mpls_lse_tc(ovs_be32 *lse, uint8_t tc);
void set_mpls_lse_label(ovs_be32 *lse, ovs_be32 label);
void set_mpls_lse_bos(ovs_be32 *lse, uint8_t bos);
ovs_be32 set_mpls_lse_values(uint8_t ttl, uint8_t tc, uint8_t bos,
ovs_be32 label);
/* Example:
*
* uint8_t mac[ETH_ADDR_LEN];
* [...]
* 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]
/* Example:
*
* char *string = "1 00:11:22:33:44:55 2";
* uint8_t mac[ETH_ADDR_LEN];
* int a, b;
*
* if (ovs_scan(string, "%d"ETH_ADDR_SCAN_FMT"%d",
* &a, ETH_ADDR_SCAN_ARGS(mac), &b)) {
* ...
* }
*/
#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_TYPE_IP 0x0800
#define ETH_TYPE_ARP 0x0806
#define ETH_TYPE_VLAN_8021Q 0x8100
#define ETH_TYPE_VLAN ETH_TYPE_VLAN_8021Q
#define ETH_TYPE_VLAN_8021AD 0x88a8
#define ETH_TYPE_IPV6 0x86dd
#define ETH_TYPE_LACP 0x8809
#define ETH_TYPE_RARP 0x8035
#define ETH_TYPE_MPLS 0x8847
#define ETH_TYPE_MPLS_MCAST 0x8848
static inline bool eth_type_mpls(ovs_be16 eth_type)
{
return eth_type == htons(ETH_TYPE_MPLS) ||
eth_type == htons(ETH_TYPE_MPLS_MCAST);
}
/* Minimum value for an Ethernet type. Values below this are IEEE 802.2 frame
* lengths. */
#define ETH_TYPE_MIN 0x600
#define ETH_HEADER_LEN 14
#define ETH_PAYLOAD_MIN 46
#define ETH_PAYLOAD_MAX 1500
#define ETH_TOTAL_MIN (ETH_HEADER_LEN + ETH_PAYLOAD_MIN)
#define ETH_TOTAL_MAX (ETH_HEADER_LEN + ETH_PAYLOAD_MAX)
#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];
2011-01-19 11:08:03 -08:00
ovs_be16 eth_type;
});
BUILD_ASSERT_DECL(ETH_HEADER_LEN == sizeof(struct eth_header));
#define LLC_DSAP_SNAP 0xaa
#define LLC_SSAP_SNAP 0xaa
#define LLC_CNTL_SNAP 3
#define LLC_HEADER_LEN 3
OVS_PACKED(
struct llc_header {
uint8_t llc_dsap;
uint8_t llc_ssap;
uint8_t llc_cntl;
});
BUILD_ASSERT_DECL(LLC_HEADER_LEN == sizeof(struct llc_header));
#define SNAP_ORG_ETHERNET "\0\0" /* The compiler adds a null byte, so
sizeof(SNAP_ORG_ETHERNET) == 3. */
#define SNAP_HEADER_LEN 5
OVS_PACKED(
struct snap_header {
uint8_t snap_org[3];
2011-01-19 11:08:03 -08:00
ovs_be16 snap_type;
});
BUILD_ASSERT_DECL(SNAP_HEADER_LEN == sizeof(struct snap_header));
#define LLC_SNAP_HEADER_LEN (LLC_HEADER_LEN + SNAP_HEADER_LEN)
OVS_PACKED(
struct llc_snap_header {
struct llc_header llc;
struct snap_header snap;
});
BUILD_ASSERT_DECL(LLC_SNAP_HEADER_LEN == sizeof(struct llc_snap_header));
#define VLAN_VID_MASK 0x0fff
#define VLAN_VID_SHIFT 0
#define VLAN_PCP_MASK 0xe000
#define VLAN_PCP_SHIFT 13
#define VLAN_CFI 0x1000
#define VLAN_CFI_SHIFT 12
/* Given the vlan_tci field from an 802.1Q header, in network byte order,
* returns the VLAN ID in host byte order. */
static inline uint16_t
2011-01-19 11:08:03 -08:00
vlan_tci_to_vid(ovs_be16 vlan_tci)
{
return (ntohs(vlan_tci) & VLAN_VID_MASK) >> VLAN_VID_SHIFT;
}
/* Given the vlan_tci field from an 802.1Q header, in network byte order,
* returns the priority code point (PCP) in host byte order. */
static inline int
2011-01-19 11:08:03 -08:00
vlan_tci_to_pcp(ovs_be16 vlan_tci)
{
return (ntohs(vlan_tci) & VLAN_PCP_MASK) >> VLAN_PCP_SHIFT;
}
/* Given the vlan_tci field from an 802.1Q header, in network byte order,
* returns the Canonical Format Indicator (CFI). */
static inline int
vlan_tci_to_cfi(ovs_be16 vlan_tci)
{
return (vlan_tci & htons(VLAN_CFI)) != 0;
}
#define VLAN_HEADER_LEN 4
struct vlan_header {
2011-01-19 11:08:03 -08:00
ovs_be16 vlan_tci; /* Lowest 12 bits are VLAN ID. */
ovs_be16 vlan_next_type;
};
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];
2011-01-19 11:08:03 -08:00
ovs_be16 veth_type; /* Always htons(ETH_TYPE_VLAN). */
ovs_be16 veth_tci; /* Lowest 12 bits are VLAN ID. */
ovs_be16 veth_next_type;
});
BUILD_ASSERT_DECL(VLAN_ETH_HEADER_LEN == sizeof(struct vlan_eth_header));
/* MPLS related definitions */
#define MPLS_TTL_MASK 0x000000ff
#define MPLS_TTL_SHIFT 0
#define MPLS_BOS_MASK 0x00000100
#define MPLS_BOS_SHIFT 8
#define MPLS_TC_MASK 0x00000e00
#define MPLS_TC_SHIFT 9
#define MPLS_LABEL_MASK 0xfffff000
#define MPLS_LABEL_SHIFT 12
#define MPLS_HLEN 4
struct mpls_hdr {
ovs_16aligned_be32 mpls_lse;
};
BUILD_ASSERT_DECL(MPLS_HLEN == sizeof(struct mpls_hdr));
/* Given a mpls label stack entry in network byte order
* return mpls label in host byte order */
static inline uint32_t
mpls_lse_to_label(ovs_be32 mpls_lse)
{
return (ntohl(mpls_lse) & MPLS_LABEL_MASK) >> MPLS_LABEL_SHIFT;
}
/* Given a mpls label stack entry in network byte order
* return mpls tc */
static inline uint8_t
mpls_lse_to_tc(ovs_be32 mpls_lse)
{
return (ntohl(mpls_lse) & MPLS_TC_MASK) >> MPLS_TC_SHIFT;
}
/* Given a mpls label stack entry in network byte order
* return mpls ttl */
static inline uint8_t
mpls_lse_to_ttl(ovs_be32 mpls_lse)
{
return (ntohl(mpls_lse) & MPLS_TTL_MASK) >> MPLS_TTL_SHIFT;
}
/* Set TTL in mpls lse. */
static inline void
flow_set_mpls_lse_ttl(ovs_be32 *mpls_lse, uint8_t ttl)
{
*mpls_lse &= ~htonl(MPLS_TTL_MASK);
*mpls_lse |= htonl(ttl << MPLS_TTL_SHIFT);
}
/* Given a mpls label stack entry in network byte order
* return mpls BoS bit */
static inline uint8_t
mpls_lse_to_bos(ovs_be32 mpls_lse)
{
return (mpls_lse & htonl(MPLS_BOS_MASK)) != 0;
}
#define IP_FMT "%"PRIu32".%"PRIu32".%"PRIu32".%"PRIu32
#define IP_ARGS(ip) \
ntohl(ip) >> 24, \
(ntohl(ip) >> 16) & 0xff, \
(ntohl(ip) >> 8) & 0xff, \
ntohl(ip) & 0xff
/* Example:
*
* char *string = "1 33.44.55.66 2";
* ovs_be32 ip;
* int a, b;
*
* if (ovs_scan(string, "%d"IP_SCAN_FMT"%d", &a, IP_SCAN_ARGS(&ip), &b)) {
* ...
* }
*/
#define IP_SCAN_FMT "%"SCNu8".%"SCNu8".%"SCNu8".%"SCNu8
#define IP_SCAN_ARGS(ip) \
((void) (ovs_be32) *(ip), &((uint8_t *) ip)[0]), \
&((uint8_t *) ip)[1], \
&((uint8_t *) ip)[2], \
&((uint8_t *) ip)[3]
/* Returns true if 'netmask' is a CIDR netmask, that is, if it consists of N
* high-order 1-bits and 32-N low-order 0-bits. */
static inline bool
ip_is_cidr(ovs_be32 netmask)
{
uint32_t x = ~ntohl(netmask);
return !(x & (x + 1));
}
static inline bool
ip_is_multicast(ovs_be32 ip)
{
return (ip & htonl(0xf0000000)) == htonl(0xe0000000);
}
static inline bool
ip_is_local_multicast(ovs_be32 ip)
{
return (ip & htonl(0xffffff00)) == htonl(0xe0000000);
}
int ip_count_cidr_bits(ovs_be32 netmask);
void ip_format_masked(ovs_be32 ip, ovs_be32 mask, struct ds *);
#define IP_VER(ip_ihl_ver) ((ip_ihl_ver) >> 4)
#define IP_IHL(ip_ihl_ver) ((ip_ihl_ver) & 15)
#define IP_IHL_VER(ihl, ver) (((ver) << 4) | (ihl))
#ifndef IPPROTO_SCTP
#define IPPROTO_SCTP 132
#endif
/* TOS fields. */
#define IP_ECN_NOT_ECT 0x0
#define IP_ECN_ECT_1 0x01
#define IP_ECN_ECT_0 0x02
#define IP_ECN_CE 0x03
#define IP_ECN_MASK 0x03
#define IP_DSCP_MASK 0xfc
#define IP_VERSION 4
#define IP_DONT_FRAGMENT 0x4000 /* Don't fragment. */
#define IP_MORE_FRAGMENTS 0x2000 /* More fragments. */
#define IP_FRAG_OFF_MASK 0x1fff /* Fragment offset. */
#define IP_IS_FRAGMENT(ip_frag_off) \
((ip_frag_off) & htons(IP_MORE_FRAGMENTS | IP_FRAG_OFF_MASK))
#define IP_HEADER_LEN 20
struct ip_header {
uint8_t ip_ihl_ver;
uint8_t ip_tos;
2011-01-19 11:08:03 -08:00
ovs_be16 ip_tot_len;
ovs_be16 ip_id;
ovs_be16 ip_frag_off;
uint8_t ip_ttl;
uint8_t ip_proto;
2011-01-19 11:08:03 -08:00
ovs_be16 ip_csum;
packets: Do not assume that IPv4, TCP, or ARP headers are 32-bit aligned. Ethernet headers are 14 bytes long, so when the beginning of such a header is 32-bit aligned, the following data is misaligned. The usual trick to fix that is to start the Ethernet header on an odd-numbered 16-bit boundary. That trick works OK for Open vSwitch, but there are two problems: - OVS doesn't use that trick everywhere. Maybe it should, but it's difficult to make sure that it does consistently because the CPUs most commonly used with OVS don't care about misalignment, so we only find problems when porting. - Some protocols (GRE, VXLAN) don't use that trick, so in such a case one can properly align the inner or outer L3/L4/L7 but not both. (OVS userspace doesn't directly deal with such protocols yet, so this is just future-proofing.) - OpenFlow uses the alignment trick in a few places but not all of them. This commit starts the adoption of what I hope will be a more robust way to avoid misalignment problems and the resulting bus errors on RISC architectures. Instead of trying to ensure that 32-bit quantities are always aligned, we always read them as if they were misaligned. To ensure that they are read this way, we change their types from 32-bit types to pairs of 16-bit types. (I don't know of any protocols that offset the next header by an odd number of bytes, so a 16-bit alignment assumption seems OK.) The same would be necessary for 64-bit types in protocol headers, but we don't yet have any protocol definitions with 64-bit types. IPv6 protocol headers need the same treatment, but for those we rely on structs provided by system headers, so I'll leave them for an upcoming patch. Signed-off-by: Ben Pfaff <blp@nicira.com>
2013-08-15 10:47:39 -07:00
ovs_16aligned_be32 ip_src;
ovs_16aligned_be32 ip_dst;
};
BUILD_ASSERT_DECL(IP_HEADER_LEN == sizeof(struct ip_header));
#define ICMP_HEADER_LEN 8
struct icmp_header {
uint8_t icmp_type;
uint8_t icmp_code;
2011-01-19 11:08:03 -08:00
ovs_be16 icmp_csum;
union {
struct {
ovs_be16 id;
ovs_be16 seq;
} echo;
struct {
ovs_be16 empty;
ovs_be16 mtu;
} frag;
packets: Do not assume that IPv4, TCP, or ARP headers are 32-bit aligned. Ethernet headers are 14 bytes long, so when the beginning of such a header is 32-bit aligned, the following data is misaligned. The usual trick to fix that is to start the Ethernet header on an odd-numbered 16-bit boundary. That trick works OK for Open vSwitch, but there are two problems: - OVS doesn't use that trick everywhere. Maybe it should, but it's difficult to make sure that it does consistently because the CPUs most commonly used with OVS don't care about misalignment, so we only find problems when porting. - Some protocols (GRE, VXLAN) don't use that trick, so in such a case one can properly align the inner or outer L3/L4/L7 but not both. (OVS userspace doesn't directly deal with such protocols yet, so this is just future-proofing.) - OpenFlow uses the alignment trick in a few places but not all of them. This commit starts the adoption of what I hope will be a more robust way to avoid misalignment problems and the resulting bus errors on RISC architectures. Instead of trying to ensure that 32-bit quantities are always aligned, we always read them as if they were misaligned. To ensure that they are read this way, we change their types from 32-bit types to pairs of 16-bit types. (I don't know of any protocols that offset the next header by an odd number of bytes, so a 16-bit alignment assumption seems OK.) The same would be necessary for 64-bit types in protocol headers, but we don't yet have any protocol definitions with 64-bit types. IPv6 protocol headers need the same treatment, but for those we rely on structs provided by system headers, so I'll leave them for an upcoming patch. Signed-off-by: Ben Pfaff <blp@nicira.com>
2013-08-15 10:47:39 -07:00
ovs_16aligned_be32 gateway;
} icmp_fields;
uint8_t icmp_data[0];
};
BUILD_ASSERT_DECL(ICMP_HEADER_LEN == sizeof(struct icmp_header));
#define IGMP_HEADER_LEN 8
struct igmp_header {
uint8_t igmp_type;
uint8_t igmp_code;
ovs_be16 igmp_csum;
ovs_16aligned_be32 group;
};
BUILD_ASSERT_DECL(IGMP_HEADER_LEN == sizeof(struct igmp_header));
#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 */
#define SCTP_HEADER_LEN 12
struct sctp_header {
ovs_be16 sctp_src;
ovs_be16 sctp_dst;
ovs_16aligned_be32 sctp_vtag;
ovs_16aligned_be32 sctp_csum;
};
BUILD_ASSERT_DECL(SCTP_HEADER_LEN == sizeof(struct sctp_header));
#define UDP_HEADER_LEN 8
struct udp_header {
2011-01-19 11:08:03 -08:00
ovs_be16 udp_src;
ovs_be16 udp_dst;
ovs_be16 udp_len;
ovs_be16 udp_csum;
};
BUILD_ASSERT_DECL(UDP_HEADER_LEN == sizeof(struct udp_header));
#define TCP_FIN 0x001
#define TCP_SYN 0x002
#define TCP_RST 0x004
#define TCP_PSH 0x008
#define TCP_ACK 0x010
#define TCP_URG 0x020
#define TCP_ECE 0x040
#define TCP_CWR 0x080
#define TCP_NS 0x100
#define TCP_CTL(flags, offset) (htons((flags) | ((offset) << 12)))
#define TCP_FLAGS(tcp_ctl) (ntohs(tcp_ctl) & 0x0fff)
#define TCP_FLAGS_BE16(tcp_ctl) ((tcp_ctl) & htons(0x0fff))
#define TCP_OFFSET(tcp_ctl) (ntohs(tcp_ctl) >> 12)
#define TCP_HEADER_LEN 20
struct tcp_header {
2011-01-19 11:08:03 -08:00
ovs_be16 tcp_src;
ovs_be16 tcp_dst;
packets: Do not assume that IPv4, TCP, or ARP headers are 32-bit aligned. Ethernet headers are 14 bytes long, so when the beginning of such a header is 32-bit aligned, the following data is misaligned. The usual trick to fix that is to start the Ethernet header on an odd-numbered 16-bit boundary. That trick works OK for Open vSwitch, but there are two problems: - OVS doesn't use that trick everywhere. Maybe it should, but it's difficult to make sure that it does consistently because the CPUs most commonly used with OVS don't care about misalignment, so we only find problems when porting. - Some protocols (GRE, VXLAN) don't use that trick, so in such a case one can properly align the inner or outer L3/L4/L7 but not both. (OVS userspace doesn't directly deal with such protocols yet, so this is just future-proofing.) - OpenFlow uses the alignment trick in a few places but not all of them. This commit starts the adoption of what I hope will be a more robust way to avoid misalignment problems and the resulting bus errors on RISC architectures. Instead of trying to ensure that 32-bit quantities are always aligned, we always read them as if they were misaligned. To ensure that they are read this way, we change their types from 32-bit types to pairs of 16-bit types. (I don't know of any protocols that offset the next header by an odd number of bytes, so a 16-bit alignment assumption seems OK.) The same would be necessary for 64-bit types in protocol headers, but we don't yet have any protocol definitions with 64-bit types. IPv6 protocol headers need the same treatment, but for those we rely on structs provided by system headers, so I'll leave them for an upcoming patch. Signed-off-by: Ben Pfaff <blp@nicira.com>
2013-08-15 10:47:39 -07:00
ovs_16aligned_be32 tcp_seq;
ovs_16aligned_be32 tcp_ack;
2011-01-19 11:08:03 -08:00
ovs_be16 tcp_ctl;
ovs_be16 tcp_winsz;
ovs_be16 tcp_csum;
ovs_be16 tcp_urg;
};
BUILD_ASSERT_DECL(TCP_HEADER_LEN == sizeof(struct tcp_header));
#define ARP_HRD_ETHERNET 1
#define ARP_PRO_IP 0x0800
#define ARP_OP_REQUEST 1
#define ARP_OP_REPLY 2
#define ARP_OP_RARP 3
#define ARP_ETH_HEADER_LEN 28
struct arp_eth_header {
/* Generic members. */
2011-01-19 11:08:03 -08:00
ovs_be16 ar_hrd; /* Hardware type. */
ovs_be16 ar_pro; /* Protocol type. */
uint8_t ar_hln; /* Hardware address length. */
uint8_t ar_pln; /* Protocol address length. */
2011-01-19 11:08:03 -08:00
ovs_be16 ar_op; /* Opcode. */
/* Ethernet+IPv4 specific members. */
uint8_t ar_sha[ETH_ADDR_LEN]; /* Sender hardware address. */
packets: Do not assume that IPv4, TCP, or ARP headers are 32-bit aligned. Ethernet headers are 14 bytes long, so when the beginning of such a header is 32-bit aligned, the following data is misaligned. The usual trick to fix that is to start the Ethernet header on an odd-numbered 16-bit boundary. That trick works OK for Open vSwitch, but there are two problems: - OVS doesn't use that trick everywhere. Maybe it should, but it's difficult to make sure that it does consistently because the CPUs most commonly used with OVS don't care about misalignment, so we only find problems when porting. - Some protocols (GRE, VXLAN) don't use that trick, so in such a case one can properly align the inner or outer L3/L4/L7 but not both. (OVS userspace doesn't directly deal with such protocols yet, so this is just future-proofing.) - OpenFlow uses the alignment trick in a few places but not all of them. This commit starts the adoption of what I hope will be a more robust way to avoid misalignment problems and the resulting bus errors on RISC architectures. Instead of trying to ensure that 32-bit quantities are always aligned, we always read them as if they were misaligned. To ensure that they are read this way, we change their types from 32-bit types to pairs of 16-bit types. (I don't know of any protocols that offset the next header by an odd number of bytes, so a 16-bit alignment assumption seems OK.) The same would be necessary for 64-bit types in protocol headers, but we don't yet have any protocol definitions with 64-bit types. IPv6 protocol headers need the same treatment, but for those we rely on structs provided by system headers, so I'll leave them for an upcoming patch. Signed-off-by: Ben Pfaff <blp@nicira.com>
2013-08-15 10:47:39 -07:00
ovs_16aligned_be32 ar_spa; /* Sender protocol address. */
uint8_t ar_tha[ETH_ADDR_LEN]; /* Target hardware address. */
packets: Do not assume that IPv4, TCP, or ARP headers are 32-bit aligned. Ethernet headers are 14 bytes long, so when the beginning of such a header is 32-bit aligned, the following data is misaligned. The usual trick to fix that is to start the Ethernet header on an odd-numbered 16-bit boundary. That trick works OK for Open vSwitch, but there are two problems: - OVS doesn't use that trick everywhere. Maybe it should, but it's difficult to make sure that it does consistently because the CPUs most commonly used with OVS don't care about misalignment, so we only find problems when porting. - Some protocols (GRE, VXLAN) don't use that trick, so in such a case one can properly align the inner or outer L3/L4/L7 but not both. (OVS userspace doesn't directly deal with such protocols yet, so this is just future-proofing.) - OpenFlow uses the alignment trick in a few places but not all of them. This commit starts the adoption of what I hope will be a more robust way to avoid misalignment problems and the resulting bus errors on RISC architectures. Instead of trying to ensure that 32-bit quantities are always aligned, we always read them as if they were misaligned. To ensure that they are read this way, we change their types from 32-bit types to pairs of 16-bit types. (I don't know of any protocols that offset the next header by an odd number of bytes, so a 16-bit alignment assumption seems OK.) The same would be necessary for 64-bit types in protocol headers, but we don't yet have any protocol definitions with 64-bit types. IPv6 protocol headers need the same treatment, but for those we rely on structs provided by system headers, so I'll leave them for an upcoming patch. Signed-off-by: Ben Pfaff <blp@nicira.com>
2013-08-15 10:47:39 -07:00
ovs_16aligned_be32 ar_tpa; /* Target protocol address. */
};
BUILD_ASSERT_DECL(ARP_ETH_HEADER_LEN == sizeof(struct arp_eth_header));
/* 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 {
ovs_be16 be16[8];
ovs_16aligned_be32 be32[4];
};
/* Like struct in6_hdr, but whereas that struct requires 32-bit alignment, this
* one only requires 16-bit alignment. */
struct ovs_16aligned_ip6_hdr {
union {
struct ovs_16aligned_ip6_hdrctl {
ovs_16aligned_be32 ip6_un1_flow;
ovs_be16 ip6_un1_plen;
uint8_t ip6_un1_nxt;
uint8_t ip6_un1_hlim;
} ip6_un1;
uint8_t ip6_un2_vfc;
} ip6_ctlun;
union ovs_16aligned_in6_addr ip6_src;
union ovs_16aligned_in6_addr ip6_dst;
};
/* Like struct in6_frag, but whereas that struct requires 32-bit alignment,
* this one only requires 16-bit alignment. */
struct ovs_16aligned_ip6_frag {
uint8_t ip6f_nxt;
uint8_t ip6f_reserved;
ovs_be16 ip6f_offlg;
ovs_16aligned_be32 ip6f_ident;
};
/* The IPv6 flow label is in the lower 20 bits of the first 32-bit word. */
#define IPV6_LABEL_MASK 0x000fffff
/* Example:
*
* char *string = "1 ::1 2";
* char ipv6_s[IPV6_SCAN_LEN + 1];
* struct in6_addr ipv6;
*
* if (ovs_scan(string, "%d"IPV6_SCAN_FMT"%d", &a, ipv6_s, &b)
* && inet_pton(AF_INET6, ipv6_s, &ipv6) == 1) {
* ...
* }
*/
#define IPV6_SCAN_FMT "%46[0123456789abcdefABCDEF:.]"
#define IPV6_SCAN_LEN 46
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 } } }
static inline bool ipv6_addr_equals(const struct in6_addr *a,
const struct in6_addr *b)
{
#ifdef IN6_ARE_ADDR_EQUAL
return IN6_ARE_ADDR_EQUAL(a, b);
#else
return !memcmp(a, b, sizeof(*a));
#endif
}
static inline bool ipv6_mask_is_any(const struct in6_addr *mask) {
return ipv6_addr_equals(mask, &in6addr_any);
}
static inline bool ipv6_mask_is_exact(const struct in6_addr *mask) {
return ipv6_addr_equals(mask, &in6addr_exact);
}
static inline bool dl_type_is_ip_any(ovs_be16 dl_type)
{
return dl_type == htons(ETH_TYPE_IP)
|| dl_type == htons(ETH_TYPE_IPV6);
}
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);
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 ofpbuf *, const uint8_t eth_dst[ETH_ADDR_LEN],
const uint8_t eth_src[ETH_ADDR_LEN], uint16_t eth_type,
size_t size);
void *snap_compose(struct ofpbuf *, const uint8_t eth_dst[ETH_ADDR_LEN],
const uint8_t eth_src[ETH_ADDR_LEN],
unsigned int oui, uint16_t snap_type, size_t size);
void packet_set_ipv4(struct ofpbuf *, ovs_be32 src, ovs_be32 dst, uint8_t tos,
uint8_t ttl);
void packet_set_ipv6(struct ofpbuf *, uint8_t proto, const ovs_be32 src[4],
const ovs_be32 dst[4], uint8_t tc,
ovs_be32 fl, uint8_t hlmit);
void packet_set_tcp_port(struct ofpbuf *, ovs_be16 src, ovs_be16 dst);
void packet_set_udp_port(struct ofpbuf *, ovs_be16 src, ovs_be16 dst);
void packet_set_sctp_port(struct ofpbuf *, ovs_be16 src, ovs_be16 dst);
void packet_format_tcp_flags(struct ds *, uint16_t);
const char *packet_tcp_flag_to_string(uint32_t flag);
#endif /* packets.h */