2 * Copyright (c) 2007-2017 Nicira, Inc.
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of version 2 of the GNU General Public
6 * License as published by the Free Software Foundation.
8 * This program is distributed in the hope that it will be useful, but
9 * WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
13 * You should have received a copy of the GNU General Public License
14 * along with this program; if not, write to the Free Software
15 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
22 #include <linux/cache.h>
23 #include <linux/kernel.h>
24 #include <linux/netlink.h>
25 #include <linux/openvswitch.h>
26 #include <linux/spinlock.h>
27 #include <linux/types.h>
28 #include <linux/rcupdate.h>
29 #include <linux/if_ether.h>
30 #include <linux/in6.h>
31 #include <linux/jiffies.h>
32 #include <linux/time.h>
33 #include <linux/flex_array.h>
34 #include <net/inet_ecn.h>
35 #include <net/ip_tunnels.h>
36 #include <net/dst_metadata.h>
40 enum sw_flow_mac_proto
{
44 #define SW_FLOW_KEY_INVALID 0x80
46 /* Store options at the end of the array if they are less than the
47 * maximum size. This allows us to get the benefits of variable length
48 * matching for small options.
50 #define TUN_METADATA_OFFSET(opt_len) \
51 (FIELD_SIZEOF(struct sw_flow_key, tun_opts) - opt_len)
52 #define TUN_METADATA_OPTS(flow_key, opt_len) \
53 ((void *)((flow_key)->tun_opts + TUN_METADATA_OFFSET(opt_len)))
55 struct ovs_tunnel_info
{
56 struct metadata_dst
*tun_dst
;
60 __be16 tpid
; /* Vlan type. Generally 802.1q or 802.1ad.*/
61 __be16 tci
; /* 0 if no VLAN, VLAN_TAG_PRESENT set otherwise. */
64 #define OVS_SW_FLOW_KEY_METADATA_SIZE \
65 (offsetof(struct sw_flow_key, recirc_id) + \
66 FIELD_SIZEOF(struct sw_flow_key, recirc_id))
69 u8 tun_opts
[IP_TUNNEL_OPTS_MAX
];
71 struct ip_tunnel_key tun_key
; /* Encapsulating tunnel key. */
73 u32 priority
; /* Packet QoS priority. */
74 u32 skb_mark
; /* SKB mark. */
75 u16 in_port
; /* Input switch port (or DP_MAX_PORTS). */
76 } __packed phy
; /* Safe when right after 'tun_key'. */
77 u8 mac_proto
; /* MAC layer protocol (e.g. Ethernet). */
78 u8 tun_proto
; /* Protocol of encapsulating tunnel. */
79 u32 ovs_flow_hash
; /* Datapath computed hash value. */
80 u32 recirc_id
; /* Recirculation ID. */
82 u8 src
[ETH_ALEN
]; /* Ethernet source address. */
83 u8 dst
[ETH_ALEN
]; /* Ethernet destination address. */
84 struct vlan_head vlan
;
85 struct vlan_head cvlan
;
86 __be16 type
; /* Ethernet frame type. */
88 /* Filling a hole of two bytes. */
90 u8 ct_orig_proto
; /* CT original direction tuple IP
95 __be32 top_lse
; /* top label stack entry */
98 u8 proto
; /* IP protocol or lower 8 bits of ARP opcode. */
100 u8 ttl
; /* IP TTL/hop limit. */
101 u8 frag
; /* One of OVS_FRAG_TYPE_*. */
104 u16 ct_zone
; /* Conntrack zone. */
106 __be16 src
; /* TCP/UDP/SCTP source port. */
107 __be16 dst
; /* TCP/UDP/SCTP destination port. */
108 __be16 flags
; /* TCP flags. */
113 __be32 src
; /* IP source address. */
114 __be32 dst
; /* IP destination address. */
120 } ct_orig
; /* Conntrack original direction fields. */
122 u8 sha
[ETH_ALEN
]; /* ARP source hardware address. */
123 u8 tha
[ETH_ALEN
]; /* ARP target hardware address. */
129 struct in6_addr src
; /* IPv6 source address. */
130 struct in6_addr dst
; /* IPv6 destination address. */
132 __be32 label
; /* IPv6 flow label. */
137 } ct_orig
; /* Conntrack original direction fields. */
139 struct in6_addr target
; /* ND target address. */
140 u8 sll
[ETH_ALEN
]; /* ND source link layer address. */
141 u8 tll
[ETH_ALEN
]; /* ND target link layer address. */
147 /* Connection tracking fields not packed above. */
149 __be16 src
; /* CT orig tuple tp src port. */
150 __be16 dst
; /* CT orig tuple tp dst port. */
153 struct ovs_key_ct_labels labels
;
156 } __aligned(BITS_PER_LONG
/8); /* Ensure that we can do comparisons as longs. */
158 static inline bool sw_flow_key_is_nd(const struct sw_flow_key
*key
)
160 return key
->eth
.type
== htons(ETH_P_IPV6
) &&
161 key
->ip
.proto
== NEXTHDR_ICMP
&&
163 (key
->tp
.src
== htons(NDISC_NEIGHBOUR_SOLICITATION
) ||
164 key
->tp
.src
== htons(NDISC_NEIGHBOUR_ADVERTISEMENT
));
167 struct sw_flow_key_range
{
168 unsigned short int start
;
169 unsigned short int end
;
172 struct sw_flow_mask
{
175 struct list_head list
;
176 struct sw_flow_key_range range
;
177 struct sw_flow_key key
;
180 struct sw_flow_match
{
181 struct sw_flow_key
*key
;
182 struct sw_flow_key_range range
;
183 struct sw_flow_mask
*mask
;
186 #define MAX_UFID_LENGTH 16 /* 128 bits */
191 u32 ufid
[MAX_UFID_LENGTH
/ 4];
192 struct sw_flow_key
*unmasked_key
;
196 struct sw_flow_actions
{
198 size_t orig_len
; /* From flow_cmd_new netlink actions size */
200 struct nlattr actions
[];
204 u64 packet_count
; /* Number of packets matched. */
205 u64 byte_count
; /* Number of bytes matched. */
206 unsigned long used
; /* Last used time (in jiffies). */
207 spinlock_t lock
; /* Lock for atomic stats update. */
208 __be16 tcp_flags
; /* Union of seen TCP flags. */
214 struct hlist_node node
[2];
216 } flow_table
, ufid_table
;
217 int stats_last_writer
; /* CPU id of the last writer on
220 struct sw_flow_key key
;
221 struct sw_flow_id id
;
222 struct sw_flow_mask
*mask
;
223 struct sw_flow_actions __rcu
*sf_acts
;
224 struct flow_stats __rcu
*stats
[]; /* One for each CPU. First one
225 * is allocated at flow creation time,
226 * the rest are allocated on demand
227 * while holding the 'stats[0].lock'.
231 struct arp_eth_header
{
232 __be16 ar_hrd
; /* format of hardware address */
233 __be16 ar_pro
; /* format of protocol address */
234 unsigned char ar_hln
; /* length of hardware address */
235 unsigned char ar_pln
; /* length of protocol address */
236 __be16 ar_op
; /* ARP opcode (command) */
238 /* Ethernet+IPv4 specific members. */
239 unsigned char ar_sha
[ETH_ALEN
]; /* sender hardware address */
240 unsigned char ar_sip
[4]; /* sender IP address */
241 unsigned char ar_tha
[ETH_ALEN
]; /* target hardware address */
242 unsigned char ar_tip
[4]; /* target IP address */
245 static inline u8
ovs_key_mac_proto(const struct sw_flow_key
*key
)
247 return key
->mac_proto
& ~SW_FLOW_KEY_INVALID
;
250 static inline u16
__ovs_mac_header_len(u8 mac_proto
)
252 return mac_proto
== MAC_PROTO_ETHERNET
? ETH_HLEN
: 0;
255 static inline u16
ovs_mac_header_len(const struct sw_flow_key
*key
)
257 return __ovs_mac_header_len(ovs_key_mac_proto(key
));
260 static inline bool ovs_identifier_is_ufid(const struct sw_flow_id
*sfid
)
262 return sfid
->ufid_len
;
265 static inline bool ovs_identifier_is_key(const struct sw_flow_id
*sfid
)
267 return !ovs_identifier_is_ufid(sfid
);
270 void ovs_flow_stats_update(struct sw_flow
*, __be16 tcp_flags
,
271 const struct sk_buff
*);
272 void ovs_flow_stats_get(const struct sw_flow
*, struct ovs_flow_stats
*,
273 unsigned long *used
, __be16
*tcp_flags
);
274 void ovs_flow_stats_clear(struct sw_flow
*);
275 u64
ovs_flow_used_time(unsigned long flow_jiffies
);
277 int ovs_flow_key_update(struct sk_buff
*skb
, struct sw_flow_key
*key
);
278 int ovs_flow_key_extract(const struct ip_tunnel_info
*tun_info
,
280 struct sw_flow_key
*key
);
281 /* Extract key from packet coming from userspace. */
282 int ovs_flow_key_extract_userspace(struct net
*net
, const struct nlattr
*attr
,
284 struct sw_flow_key
*key
, bool log
);