X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=datapath%2Fflow.c;h=29d306230c70584880531badeea763157e400491;hb=78f1e5c7d39a27d86677bb7c301824f8a8c3a5a8;hp=d32bbdea97ad5aec054b8e9c557203bd94d17a31;hpb=acd051f1761569205827dc9b037e15568a8d59f8;p=sliver-openvswitch.git diff --git a/datapath/flow.c b/datapath/flow.c index d32bbdea9..29d306230 100644 --- a/datapath/flow.c +++ b/datapath/flow.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2007-2011 Nicira Networks. + * Copyright (c) 2007-2013 Nicira, Inc. * * This program is free software; you can redistribute it and/or * modify it under the terms of version 2 of the GNU General Public @@ -32,7 +32,6 @@ #include #include #include -#include #include #include #include @@ -48,6 +47,190 @@ static struct kmem_cache *flow_cache; +static void ovs_sw_flow_mask_set(struct sw_flow_mask *mask, + struct sw_flow_key_range *range, u8 val); + +static void update_range__(struct sw_flow_match *match, + size_t offset, size_t size, bool is_mask) +{ + struct sw_flow_key_range *range = NULL; + size_t start = offset; + size_t end = offset + size; + + if (!is_mask) + range = &match->range; + else if (match->mask) + range = &match->mask->range; + + if (!range) + return; + + if (range->start == range->end) { + range->start = start; + range->end = end; + return; + } + + if (range->start > start) + range->start = start; + + if (range->end < end) + range->end = end; +} + +#define SW_FLOW_KEY_PUT(match, field, value, is_mask) \ + do { \ + update_range__(match, offsetof(struct sw_flow_key, field), \ + sizeof((match)->key->field), is_mask); \ + if (is_mask) { \ + if ((match)->mask) \ + (match)->mask->key.field = value; \ + } else { \ + (match)->key->field = value; \ + } \ + } while (0) + +#define SW_FLOW_KEY_MEMCPY(match, field, value_p, len, is_mask) \ + do { \ + update_range__(match, offsetof(struct sw_flow_key, field), \ + len, is_mask); \ + if (is_mask) { \ + if ((match)->mask) \ + memcpy(&(match)->mask->key.field, value_p, len);\ + } else { \ + memcpy(&(match)->key->field, value_p, len); \ + } \ + } while (0) + +void ovs_match_init(struct sw_flow_match *match, + struct sw_flow_key *key, + struct sw_flow_mask *mask) +{ + memset(match, 0, sizeof(*match)); + match->key = key; + match->mask = mask; + + memset(key, 0, sizeof(*key)); + + if (mask) { + memset(&mask->key, 0, sizeof(mask->key)); + mask->range.start = mask->range.end = 0; + } +} + +static bool ovs_match_validate(const struct sw_flow_match *match, + u64 key_attrs, u64 mask_attrs) +{ + u64 key_expected = 1ULL << OVS_KEY_ATTR_ETHERNET; + u64 mask_allowed = key_attrs; /* At most allow all key attributes */ + + /* The following mask attributes allowed only if they + * pass the validation tests. */ + mask_allowed &= ~((1ULL << OVS_KEY_ATTR_IPV4) + | (1ULL << OVS_KEY_ATTR_IPV6) + | (1ULL << OVS_KEY_ATTR_TCP) + | (1ULL << OVS_KEY_ATTR_UDP) + | (1ULL << OVS_KEY_ATTR_ICMP) + | (1ULL << OVS_KEY_ATTR_ICMPV6) + | (1ULL << OVS_KEY_ATTR_ARP) + | (1ULL << OVS_KEY_ATTR_ND)); + + /* Tunnel mask is always allowed. */ + mask_allowed |= (1ULL << OVS_KEY_ATTR_TUNNEL); + + if (match->key->phy.in_port == DP_MAX_PORTS && + match->mask && (match->mask->key.phy.in_port == 0xffff)) + mask_allowed |= (1ULL << OVS_KEY_ATTR_IN_PORT); + + if (match->key->eth.type == htons(ETH_P_802_2) && + match->mask && (match->mask->key.eth.type == htons(0xffff))) + mask_allowed |= (1ULL << OVS_KEY_ATTR_ETHERTYPE); + + /* Check key attributes. */ + if (match->key->eth.type == htons(ETH_P_ARP) + || match->key->eth.type == htons(ETH_P_RARP)) { + key_expected |= 1ULL << OVS_KEY_ATTR_ARP; + if (match->mask && (match->mask->key.eth.type == htons(0xffff))) + mask_allowed |= 1ULL << OVS_KEY_ATTR_ARP; + } + + if (match->key->eth.type == htons(ETH_P_IP)) { + key_expected |= 1ULL << OVS_KEY_ATTR_IPV4; + if (match->mask && (match->mask->key.eth.type == htons(0xffff))) + mask_allowed |= 1ULL << OVS_KEY_ATTR_IPV4; + + if (match->key->ip.frag != OVS_FRAG_TYPE_LATER) { + if (match->key->ip.proto == IPPROTO_UDP) { + key_expected |= 1ULL << OVS_KEY_ATTR_UDP; + if (match->mask && (match->mask->key.ip.proto == 0xff)) + mask_allowed |= 1ULL << OVS_KEY_ATTR_UDP; + } + + if (match->key->ip.proto == IPPROTO_TCP) { + key_expected |= 1ULL << OVS_KEY_ATTR_TCP; + if (match->mask && (match->mask->key.ip.proto == 0xff)) + mask_allowed |= 1ULL << OVS_KEY_ATTR_TCP; + } + + if (match->key->ip.proto == IPPROTO_ICMP) { + key_expected |= 1ULL << OVS_KEY_ATTR_ICMP; + if (match->mask && (match->mask->key.ip.proto == 0xff)) + mask_allowed |= 1ULL << OVS_KEY_ATTR_ICMP; + } + } + } + + if (match->key->eth.type == htons(ETH_P_IPV6)) { + key_expected |= 1ULL << OVS_KEY_ATTR_IPV6; + if (match->mask && (match->mask->key.eth.type == htons(0xffff))) + mask_allowed |= 1ULL << OVS_KEY_ATTR_IPV6; + + if (match->key->ip.frag != OVS_FRAG_TYPE_LATER) { + if (match->key->ip.proto == IPPROTO_UDP) { + key_expected |= 1ULL << OVS_KEY_ATTR_UDP; + if (match->mask && (match->mask->key.ip.proto == 0xff)) + mask_allowed |= 1ULL << OVS_KEY_ATTR_UDP; + } + + if (match->key->ip.proto == IPPROTO_TCP) { + key_expected |= 1ULL << OVS_KEY_ATTR_TCP; + if (match->mask && (match->mask->key.ip.proto == 0xff)) + mask_allowed |= 1ULL << OVS_KEY_ATTR_TCP; + } + + if (match->key->ip.proto == IPPROTO_ICMPV6) { + key_expected |= 1ULL << OVS_KEY_ATTR_ICMPV6; + if (match->mask && (match->mask->key.ip.proto == 0xff)) + mask_allowed |= 1ULL << OVS_KEY_ATTR_ICMPV6; + + if (match->key->ipv6.tp.src == + htons(NDISC_NEIGHBOUR_SOLICITATION) || + match->key->ipv6.tp.src == htons(NDISC_NEIGHBOUR_ADVERTISEMENT)) { + key_expected |= 1ULL << OVS_KEY_ATTR_ND; + if (match->mask && (match->mask->key.ipv6.tp.src == htons(0xffff))) + mask_allowed |= 1ULL << OVS_KEY_ATTR_ND; + } + } + } + } + + if ((key_attrs & key_expected) != key_expected) { + /* Key attributes check failed. */ + OVS_NLERR("Missing expected key attributes (key_attrs=%llx, expected=%llx).\n", + key_attrs, key_expected); + return false; + } + + if ((mask_attrs & mask_allowed) != mask_attrs) { + /* Mask attributes check failed. */ + OVS_NLERR("Contain more than allowed mask fields (mask_attrs=%llx, mask_allowed=%llx).\n", + mask_attrs, mask_allowed); + return false; + } + + return true; +} + static int check_header(struct sk_buff *skb, int len) { if (unlikely(skb->len < len)) @@ -123,12 +306,7 @@ u64 ovs_flow_used_time(unsigned long flow_jiffies) return cur_ms - idle_ms; } -#define SW_FLOW_KEY_OFFSET(field) \ - (offsetof(struct sw_flow_key, field) + \ - FIELD_SIZEOF(struct sw_flow_key, field)) - -static int parse_ipv6hdr(struct sk_buff *skb, struct sw_flow_key *key, - int *key_lenp) +static int parse_ipv6hdr(struct sk_buff *skb, struct sw_flow_key *key) { unsigned int nh_ofs = skb_network_offset(skb); unsigned int nh_len; @@ -138,8 +316,6 @@ static int parse_ipv6hdr(struct sk_buff *skb, struct sw_flow_key *key, __be16 frag_off; int err; - *key_lenp = SW_FLOW_KEY_OFFSET(ipv6.label); - err = check_header(skb, nh_ofs + sizeof(*nh)); if (unlikely(err)) return err; @@ -178,6 +354,21 @@ static bool icmp6hdr_ok(struct sk_buff *skb) sizeof(struct icmp6hdr)); } +void ovs_flow_key_mask(struct sw_flow_key *dst, const struct sw_flow_key *src, + const struct sw_flow_mask *mask) +{ + u8 *m = (u8 *)&mask->key + mask->range.start; + u8 *s = (u8 *)src + mask->range.start; + u8 *d = (u8 *)dst + mask->range.start; + int i; + + memset(dst, 0, sizeof(*dst)); + for (i = 0; i < ovs_sw_flow_mask_size_roundup(mask); i++) { + *d = *s & *m; + d++, s++, m++; + } +} + #define TCP_FLAGS_OFFSET 13 #define TCP_FLAG_MASK 0x3f @@ -185,8 +376,10 @@ void ovs_flow_used(struct sw_flow *flow, struct sk_buff *skb) { u8 tcp_flags = 0; - if (flow->key.eth.type == htons(ETH_P_IP) && - flow->key.ip.proto == IPPROTO_TCP) { + if ((flow->key.eth.type == htons(ETH_P_IP) || + flow->key.eth.type == htons(ETH_P_IPV6)) && + flow->key.ip.proto == IPPROTO_TCP && + likely(skb->len >= skb_transport_offset(skb) + sizeof(struct tcphdr))) { u8 *tcp = (u8 *)tcp_hdr(skb); tcp_flags = *(tcp + TCP_FLAGS_OFFSET) & TCP_FLAG_MASK; } @@ -199,23 +392,18 @@ void ovs_flow_used(struct sw_flow *flow, struct sk_buff *skb) spin_unlock(&flow->lock); } -struct sw_flow_actions *ovs_flow_actions_alloc(const struct nlattr *actions) +struct sw_flow_actions *ovs_flow_actions_alloc(int size) { - int actions_len = nla_len(actions); struct sw_flow_actions *sfa; - /* At least DP_MAX_PORTS actions are required to be able to flood a - * packet to every port. Factor of 2 allows for setting VLAN tags, - * etc. */ - if (actions_len > 2 * DP_MAX_PORTS * nla_total_size(4)) + if (size > MAX_ACTIONS_BUFSIZE) return ERR_PTR(-EINVAL); - sfa = kmalloc(sizeof(*sfa) + actions_len, GFP_KERNEL); + sfa = kmalloc(sizeof(*sfa) + size, GFP_KERNEL); if (!sfa) return ERR_PTR(-ENOMEM); - sfa->actions_len = actions_len; - memcpy(sfa->actions, nla_data(actions), actions_len); + sfa->actions_len = 0; return sfa; } @@ -228,9 +416,8 @@ struct sw_flow *ovs_flow_alloc(void) return ERR_PTR(-ENOMEM); spin_lock_init(&flow->lock); - atomic_set(&flow->refcnt, 1); flow->sf_acts = NULL; - flow->dead = false; + flow->mask = NULL; return flow; } @@ -247,7 +434,7 @@ static struct flex_array *alloc_buckets(unsigned int n_buckets) struct flex_array *buckets; int i, err; - buckets = flex_array_alloc(sizeof(struct hlist_head *), + buckets = flex_array_alloc(sizeof(struct hlist_head), n_buckets, GFP_KERNEL); if (!buckets) return NULL; @@ -270,7 +457,7 @@ static void free_buckets(struct flex_array *buckets) flex_array_free(buckets); } -struct flow_table *ovs_flow_tbl_alloc(int new_size) +static struct flow_table *__flow_tbl_alloc(int new_size) { struct flow_table *table = kmalloc(sizeof(*table), GFP_KERNEL); @@ -288,63 +475,78 @@ struct flow_table *ovs_flow_tbl_alloc(int new_size) table->node_ver = 0; table->keep_flows = false; get_random_bytes(&table->hash_seed, sizeof(u32)); + table->mask_list = NULL; return table; } -static void flow_free(struct sw_flow *flow) -{ - flow->dead = true; - ovs_flow_put(flow); -} - -void ovs_flow_tbl_destroy(struct flow_table *table) +static void __flow_tbl_destroy(struct flow_table *table) { int i; - if (!table) - return; - if (table->keep_flows) goto skip_flows; for (i = 0; i < table->n_buckets; i++) { struct sw_flow *flow; struct hlist_head *head = flex_array_get(table->buckets, i); - struct hlist_node *node, *n; + struct hlist_node *n; int ver = table->node_ver; - hlist_for_each_entry_safe(flow, node, n, head, hash_node[ver]) { - hlist_del_rcu(&flow->hash_node[ver]); - flow_free(flow); + hlist_for_each_entry_safe(flow, n, head, hash_node[ver]) { + hlist_del(&flow->hash_node[ver]); + ovs_flow_free(flow, false); } } + BUG_ON(!list_empty(table->mask_list)); + kfree(table->mask_list); + skip_flows: free_buckets(table->buckets); kfree(table); } +struct flow_table *ovs_flow_tbl_alloc(int new_size) +{ + struct flow_table *table = __flow_tbl_alloc(new_size); + + if (!table) + return NULL; + + table->mask_list = kmalloc(sizeof(struct list_head), GFP_KERNEL); + if (!table->mask_list) { + table->keep_flows = true; + __flow_tbl_destroy(table); + return NULL; + } + INIT_LIST_HEAD(table->mask_list); + + return table; +} + static void flow_tbl_destroy_rcu_cb(struct rcu_head *rcu) { struct flow_table *table = container_of(rcu, struct flow_table, rcu); - ovs_flow_tbl_destroy(table); + __flow_tbl_destroy(table); } -void ovs_flow_tbl_deferred_destroy(struct flow_table *table) +void ovs_flow_tbl_destroy(struct flow_table *table, bool deferred) { if (!table) return; - call_rcu(&table->rcu, flow_tbl_destroy_rcu_cb); + if (deferred) + call_rcu(&table->rcu, flow_tbl_destroy_rcu_cb); + else + __flow_tbl_destroy(table); } -struct sw_flow *ovs_flow_tbl_next(struct flow_table *table, u32 *bucket, u32 *last) +struct sw_flow *ovs_flow_dump_next(struct flow_table *table, u32 *bucket, u32 *last) { struct sw_flow *flow; struct hlist_head *head; - struct hlist_node *n; int ver; int i; @@ -352,7 +554,7 @@ struct sw_flow *ovs_flow_tbl_next(struct flow_table *table, u32 *bucket, u32 *la while (*bucket < table->n_buckets) { i = 0; head = flex_array_get(table->buckets, *bucket); - hlist_for_each_entry_rcu(flow, n, head, hash_node[ver]) { + hlist_for_each_entry_rcu(flow, head, hash_node[ver]) { if (i < *last) { i++; continue; @@ -367,6 +569,16 @@ struct sw_flow *ovs_flow_tbl_next(struct flow_table *table, u32 *bucket, u32 *la return NULL; } +static void __tbl_insert(struct flow_table *table, struct sw_flow *flow) +{ + struct hlist_head *head; + + head = find_bucket(table, flow->hash); + hlist_add_head_rcu(&flow->hash_node[table->node_ver], head); + + table->count++; +} + static void flow_table_copy_flows(struct flow_table *old, struct flow_table *new) { int old_ver; @@ -379,13 +591,14 @@ static void flow_table_copy_flows(struct flow_table *old, struct flow_table *new for (i = 0; i < old->n_buckets; i++) { struct sw_flow *flow; struct hlist_head *head; - struct hlist_node *n; head = flex_array_get(old->buckets, i); - hlist_for_each_entry(flow, n, head, hash_node[old_ver]) - ovs_flow_tbl_insert(new, flow); + hlist_for_each_entry(flow, head, hash_node[old_ver]) + __tbl_insert(new, flow); } + + new->mask_list = old->mask_list; old->keep_flows = true; } @@ -393,7 +606,7 @@ static struct flow_table *__flow_tbl_rehash(struct flow_table *table, int n_buck { struct flow_table *new_table; - new_table = ovs_flow_tbl_alloc(n_buckets); + new_table = __flow_tbl_alloc(n_buckets); if (!new_table) return ERR_PTR(-ENOMEM); @@ -412,36 +625,30 @@ struct flow_table *ovs_flow_tbl_expand(struct flow_table *table) return __flow_tbl_rehash(table, table->n_buckets * 2); } -/* RCU callback used by ovs_flow_deferred_free. */ -static void rcu_free_flow_callback(struct rcu_head *rcu) +static void __flow_free(struct sw_flow *flow) { - struct sw_flow *flow = container_of(rcu, struct sw_flow, rcu); - - flow->dead = true; - ovs_flow_put(flow); + kfree((struct sf_flow_acts __force *)flow->sf_acts); + kmem_cache_free(flow_cache, flow); } -/* Schedules 'flow' to be freed after the next RCU grace period. - * The caller must hold rcu_read_lock for this to be sensible. */ -void ovs_flow_deferred_free(struct sw_flow *flow) +static void rcu_free_flow_callback(struct rcu_head *rcu) { - call_rcu(&flow->rcu, rcu_free_flow_callback); -} + struct sw_flow *flow = container_of(rcu, struct sw_flow, rcu); -void ovs_flow_hold(struct sw_flow *flow) -{ - atomic_inc(&flow->refcnt); + __flow_free(flow); } -void ovs_flow_put(struct sw_flow *flow) +void ovs_flow_free(struct sw_flow *flow, bool deferred) { - if (unlikely(!flow)) + if (!flow) return; - if (atomic_dec_and_test(&flow->refcnt)) { - kfree((struct sf_flow_acts __force *)flow->sf_acts); - kmem_cache_free(flow_cache, flow); - } + ovs_sw_flow_mask_del_ref(flow->mask, deferred); + + if (deferred) + call_rcu(&flow->rcu, rcu_free_flow_callback); + else + __flow_free(flow); } /* RCU callback used by ovs_flow_deferred_free_acts. */ @@ -496,7 +703,7 @@ static __be16 parse_ethertype(struct sk_buff *skb) proto = *(__be16 *) skb->data; __skb_pull(skb, sizeof(__be16)); - if (ntohs(proto) >= 1536) + if (ntohs(proto) >= ETH_P_802_3_MIN) return proto; if (skb->len < sizeof(struct llc_snap_hdr)) @@ -512,22 +719,23 @@ static __be16 parse_ethertype(struct sk_buff *skb) return htons(ETH_P_802_2); __skb_pull(skb, sizeof(struct llc_snap_hdr)); - return llc->ethertype; + + if (ntohs(llc->ethertype) >= ETH_P_802_3_MIN) + return llc->ethertype; + + return htons(ETH_P_802_2); } static int parse_icmpv6(struct sk_buff *skb, struct sw_flow_key *key, - int *key_lenp, int nh_len) + int nh_len) { struct icmp6hdr *icmp = icmp6_hdr(skb); - int error = 0; - int key_len; /* The ICMPv6 type and code fields use the 16-bit transport port * fields, so we need to store them in 16-bit network byte order. */ key->ipv6.tp.src = htons(icmp->icmp6_type); key->ipv6.tp.dst = htons(icmp->icmp6_code); - key_len = SW_FLOW_KEY_OFFSET(ipv6.tp); if (icmp->icmp6_code == 0 && (icmp->icmp6_type == NDISC_NEIGHBOUR_SOLICITATION || @@ -536,21 +744,17 @@ static int parse_icmpv6(struct sk_buff *skb, struct sw_flow_key *key, struct nd_msg *nd; int offset; - key_len = SW_FLOW_KEY_OFFSET(ipv6.nd); - /* In order to process neighbor discovery options, we need the * entire packet. */ if (unlikely(icmp_len < sizeof(*nd))) - goto out; - if (unlikely(skb_linearize(skb))) { - error = -ENOMEM; - goto out; - } + return 0; + + if (unlikely(skb_linearize(skb))) + return -ENOMEM; nd = (struct nd_msg *)skb_transport_header(skb); key->ipv6.nd.target = nd->target; - key_len = SW_FLOW_KEY_OFFSET(ipv6.nd); icmp_len -= sizeof(*nd); offset = 0; @@ -560,7 +764,7 @@ static int parse_icmpv6(struct sk_buff *skb, struct sw_flow_key *key, int opt_len = nd_opt->nd_opt_len * 8; if (unlikely(!opt_len || opt_len > icmp_len)) - goto invalid; + return 0; /* Store the link layer address if the appropriate * option is provided. It is considered an error if @@ -585,16 +789,14 @@ static int parse_icmpv6(struct sk_buff *skb, struct sw_flow_key *key, } } - goto out; + return 0; invalid: memset(&key->ipv6.nd.target, 0, sizeof(key->ipv6.nd.target)); memset(key->ipv6.nd.sll, 0, sizeof(key->ipv6.nd.sll)); memset(key->ipv6.nd.tll, 0, sizeof(key->ipv6.nd.tll)); -out: - *key_lenp = key_len; - return error; + return 0; } /** @@ -616,23 +818,23 @@ out: * - skb->network_header: just past the Ethernet header, or just past the * VLAN header, to the first byte of the Ethernet payload. * - * - skb->transport_header: If key->dl_type is ETH_P_IP or ETH_P_IPV6 + * - skb->transport_header: If key->eth.type is ETH_P_IP or ETH_P_IPV6 * on output, then just past the IP header, if one is present and * of a correct length, otherwise the same as skb->network_header. - * For other key->dl_type values it is left untouched. + * For other key->eth.type values it is left untouched. */ -int ovs_flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key, - int *key_lenp) +int ovs_flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key) { - int error = 0; - int key_len = SW_FLOW_KEY_OFFSET(eth); + int error; struct ethhdr *eth; memset(key, 0, sizeof(*key)); key->phy.priority = skb->priority; - key->phy.tun_id = OVS_CB(skb)->tun_id; + if (OVS_CB(skb)->tun_key) + memcpy(&key->tun_key, OVS_CB(skb)->tun_key, sizeof(key->tun_key)); key->phy.in_port = in_port; + key->phy.skb_mark = skb_get_mark(skb); skb_reset_mac_header(skb); @@ -644,6 +846,8 @@ int ovs_flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key, memcpy(key->eth.dst, eth->h_dest, ETH_ALEN); __skb_pull(skb, 2 * ETH_ALEN); + /* We are going to push all headers that we pull, so no need to + * update skb->csum here. */ if (vlan_tx_tag_present(skb)) key->eth.tci = htons(vlan_get_tci(skb)); @@ -663,15 +867,13 @@ int ovs_flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key, struct iphdr *nh; __be16 offset; - key_len = SW_FLOW_KEY_OFFSET(ipv4.addr); - error = check_iphdr(skb); if (unlikely(error)) { if (error == -EINVAL) { skb->transport_header = skb->network_header; error = 0; } - goto out; + return error; } nh = ip_hdr(skb); @@ -685,7 +887,7 @@ int ovs_flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key, offset = nh->frag_off & htons(IP_OFFSET); if (offset) { key->ip.frag = OVS_FRAG_TYPE_LATER; - goto out; + return 0; } if (nh->frag_off & htons(IP_MF) || skb_shinfo(skb)->gso_type & SKB_GSO_UDP) @@ -693,21 +895,18 @@ int ovs_flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key, /* Transport layer. */ if (key->ip.proto == IPPROTO_TCP) { - key_len = SW_FLOW_KEY_OFFSET(ipv4.tp); if (tcphdr_ok(skb)) { struct tcphdr *tcp = tcp_hdr(skb); key->ipv4.tp.src = tcp->source; key->ipv4.tp.dst = tcp->dest; } } else if (key->ip.proto == IPPROTO_UDP) { - key_len = SW_FLOW_KEY_OFFSET(ipv4.tp); if (udphdr_ok(skb)) { struct udphdr *udp = udp_hdr(skb); key->ipv4.tp.src = udp->source; key->ipv4.tp.dst = udp->dest; } } else if (key->ip.proto == IPPROTO_ICMP) { - key_len = SW_FLOW_KEY_OFFSET(ipv4.tp); if (icmphdr_ok(skb)) { struct icmphdr *icmp = icmp_hdr(skb); /* The ICMP type and code fields use the 16-bit @@ -718,7 +917,8 @@ int ovs_flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key, } } - } else if (key->eth.type == htons(ETH_P_ARP) && arphdr_ok(skb)) { + } else if ((key->eth.type == htons(ETH_P_ARP) || + key->eth.type == htons(ETH_P_RARP)) && arphdr_ok(skb)) { struct arp_eth_header *arp; arp = (struct arp_eth_header *)skb_network_header(skb); @@ -731,103 +931,162 @@ int ovs_flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key, /* We only match on the lower 8 bits of the opcode. */ if (ntohs(arp->ar_op) <= 0xff) key->ip.proto = ntohs(arp->ar_op); - - if (key->ip.proto == ARPOP_REQUEST - || key->ip.proto == ARPOP_REPLY) { - memcpy(&key->ipv4.addr.src, arp->ar_sip, sizeof(key->ipv4.addr.src)); - memcpy(&key->ipv4.addr.dst, arp->ar_tip, sizeof(key->ipv4.addr.dst)); - memcpy(key->ipv4.arp.sha, arp->ar_sha, ETH_ALEN); - memcpy(key->ipv4.arp.tha, arp->ar_tha, ETH_ALEN); - key_len = SW_FLOW_KEY_OFFSET(ipv4.arp); - } + memcpy(&key->ipv4.addr.src, arp->ar_sip, sizeof(key->ipv4.addr.src)); + memcpy(&key->ipv4.addr.dst, arp->ar_tip, sizeof(key->ipv4.addr.dst)); + memcpy(key->ipv4.arp.sha, arp->ar_sha, ETH_ALEN); + memcpy(key->ipv4.arp.tha, arp->ar_tha, ETH_ALEN); } } else if (key->eth.type == htons(ETH_P_IPV6)) { int nh_len; /* IPv6 Header + Extensions */ - nh_len = parse_ipv6hdr(skb, key, &key_len); + nh_len = parse_ipv6hdr(skb, key); if (unlikely(nh_len < 0)) { - if (nh_len == -EINVAL) + if (nh_len == -EINVAL) { skb->transport_header = skb->network_header; - else + error = 0; + } else { error = nh_len; - goto out; + } + return error; } if (key->ip.frag == OVS_FRAG_TYPE_LATER) - goto out; + return 0; if (skb_shinfo(skb)->gso_type & SKB_GSO_UDP) key->ip.frag = OVS_FRAG_TYPE_FIRST; /* Transport layer. */ if (key->ip.proto == NEXTHDR_TCP) { - key_len = SW_FLOW_KEY_OFFSET(ipv6.tp); if (tcphdr_ok(skb)) { struct tcphdr *tcp = tcp_hdr(skb); key->ipv6.tp.src = tcp->source; key->ipv6.tp.dst = tcp->dest; } } else if (key->ip.proto == NEXTHDR_UDP) { - key_len = SW_FLOW_KEY_OFFSET(ipv6.tp); if (udphdr_ok(skb)) { struct udphdr *udp = udp_hdr(skb); key->ipv6.tp.src = udp->source; key->ipv6.tp.dst = udp->dest; } } else if (key->ip.proto == NEXTHDR_ICMP) { - key_len = SW_FLOW_KEY_OFFSET(ipv6.tp); if (icmp6hdr_ok(skb)) { - error = parse_icmpv6(skb, key, &key_len, nh_len); - if (error < 0) - goto out; + error = parse_icmpv6(skb, key, nh_len); + if (error) + return error; } } } -out: - *key_lenp = key_len; - return error; + return 0; +} + +static u32 ovs_flow_hash(const struct sw_flow_key *key, int key_start, int key_len) +{ + return jhash2((u32 *)((u8 *)key + key_start), + DIV_ROUND_UP(key_len - key_start, sizeof(u32)), 0); +} + +static int flow_key_start(const struct sw_flow_key *key) +{ + if (key->tun_key.ipv4_dst) + return 0; + else + return offsetof(struct sw_flow_key, phy); +} + +static bool __cmp_key(const struct sw_flow_key *key1, + const struct sw_flow_key *key2, int key_start, int key_len) +{ + return !memcmp((u8 *)key1 + key_start, + (u8 *)key2 + key_start, (key_len - key_start)); +} + +static bool __flow_cmp_key(const struct sw_flow *flow, + const struct sw_flow_key *key, int key_start, int key_len) +{ + return __cmp_key(&flow->key, key, key_start, key_len); } -u32 ovs_flow_hash(const struct sw_flow_key *key, int key_len) +static bool __flow_cmp_unmasked_key(const struct sw_flow *flow, + const struct sw_flow_key *key, int key_start, int key_len) { - return jhash2((u32 *)key, DIV_ROUND_UP(key_len, sizeof(u32)), 0); + return __cmp_key(&flow->unmasked_key, key, key_start, key_len); +} + +bool ovs_flow_cmp_unmasked_key(const struct sw_flow *flow, + const struct sw_flow_key *key, int key_len) +{ + int key_start; + key_start = flow_key_start(key); + + return __flow_cmp_unmasked_key(flow, key, key_start, key_len); + +} + +struct sw_flow *ovs_flow_lookup_unmasked_key(struct flow_table *table, + struct sw_flow_match *match) +{ + struct sw_flow_key *unmasked = match->key; + int key_len = match->range.end; + struct sw_flow *flow; + + flow = ovs_flow_lookup(table, unmasked); + if (flow && (!ovs_flow_cmp_unmasked_key(flow, unmasked, key_len))) + flow = NULL; + + return flow; } -struct sw_flow *ovs_flow_tbl_lookup(struct flow_table *table, - struct sw_flow_key *key, int key_len) +static struct sw_flow *ovs_masked_flow_lookup(struct flow_table *table, + const struct sw_flow_key *flow_key, + struct sw_flow_mask *mask) { struct sw_flow *flow; - struct hlist_node *n; struct hlist_head *head; + int key_start = mask->range.start; + int key_len = mask->range.end; u32 hash; + struct sw_flow_key masked_key; - hash = ovs_flow_hash(key, key_len); - + ovs_flow_key_mask(&masked_key, flow_key, mask); + hash = ovs_flow_hash(&masked_key, key_start, key_len); head = find_bucket(table, hash); - hlist_for_each_entry_rcu(flow, n, head, hash_node[table->node_ver]) { - - if (flow->hash == hash && - !memcmp(&flow->key, key, key_len)) { + hlist_for_each_entry_rcu(flow, head, hash_node[table->node_ver]) { + if (flow->mask == mask && + __flow_cmp_key(flow, &masked_key, key_start, key_len)) return flow; - } } return NULL; } -void ovs_flow_tbl_insert(struct flow_table *table, struct sw_flow *flow) +struct sw_flow *ovs_flow_lookup(struct flow_table *tbl, + const struct sw_flow_key *key) { - struct hlist_head *head; + struct sw_flow *flow = NULL; + struct sw_flow_mask *mask; - head = find_bucket(table, flow->hash); - hlist_add_head_rcu(&flow->hash_node[table->node_ver], head); - table->count++; + list_for_each_entry_rcu(mask, tbl->mask_list, list) { + flow = ovs_masked_flow_lookup(tbl, key, mask); + if (flow) /* Found */ + break; + } + + return flow; } -void ovs_flow_tbl_remove(struct flow_table *table, struct sw_flow *flow) + +void ovs_flow_insert(struct flow_table *table, struct sw_flow *flow) { + flow->hash = ovs_flow_hash(&flow->key, flow->mask->range.start, + flow->mask->range.end); + __tbl_insert(table, flow); +} + +void ovs_flow_remove(struct flow_table *table, struct sw_flow *flow) +{ + BUG_ON(table->count == 0); hlist_del_rcu(&flow->hash_node[table->node_ver]); table->count--; - BUG_ON(table->count < 0); } /* The size of the argument for each %OVS_KEY_ATTR_* Netlink attribute. */ @@ -835,6 +1094,7 @@ const int ovs_key_lens[OVS_KEY_ATTR_MAX + 1] = { [OVS_KEY_ATTR_ENCAP] = -1, [OVS_KEY_ATTR_PRIORITY] = sizeof(u32), [OVS_KEY_ATTR_IN_PORT] = sizeof(u32), + [OVS_KEY_ATTR_SKB_MARK] = sizeof(u32), [OVS_KEY_ATTR_ETHERNET] = sizeof(struct ovs_key_ethernet), [OVS_KEY_ATTR_VLAN] = sizeof(__be16), [OVS_KEY_ATTR_ETHERTYPE] = sizeof(__be16), @@ -846,317 +1106,555 @@ const int ovs_key_lens[OVS_KEY_ATTR_MAX + 1] = { [OVS_KEY_ATTR_ICMPV6] = sizeof(struct ovs_key_icmpv6), [OVS_KEY_ATTR_ARP] = sizeof(struct ovs_key_arp), [OVS_KEY_ATTR_ND] = sizeof(struct ovs_key_nd), - - /* Not upstream. */ - [OVS_KEY_ATTR_TUN_ID] = sizeof(__be64), + [OVS_KEY_ATTR_TUNNEL] = -1, }; -static int ipv4_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_len, - const struct nlattr *a[], u64 *attrs) +static bool is_all_zero(const u8 *fp, size_t size) { - const struct ovs_key_icmp *icmp_key; - const struct ovs_key_tcp *tcp_key; - const struct ovs_key_udp *udp_key; - - switch (swkey->ip.proto) { - case IPPROTO_TCP: - if (!(*attrs & (1 << OVS_KEY_ATTR_TCP))) - return -EINVAL; - *attrs &= ~(1 << OVS_KEY_ATTR_TCP); - - *key_len = SW_FLOW_KEY_OFFSET(ipv4.tp); - tcp_key = nla_data(a[OVS_KEY_ATTR_TCP]); - swkey->ipv4.tp.src = tcp_key->tcp_src; - swkey->ipv4.tp.dst = tcp_key->tcp_dst; - break; - - case IPPROTO_UDP: - if (!(*attrs & (1 << OVS_KEY_ATTR_UDP))) - return -EINVAL; - *attrs &= ~(1 << OVS_KEY_ATTR_UDP); - - *key_len = SW_FLOW_KEY_OFFSET(ipv4.tp); - udp_key = nla_data(a[OVS_KEY_ATTR_UDP]); - swkey->ipv4.tp.src = udp_key->udp_src; - swkey->ipv4.tp.dst = udp_key->udp_dst; - break; + int i; - case IPPROTO_ICMP: - if (!(*attrs & (1 << OVS_KEY_ATTR_ICMP))) - return -EINVAL; - *attrs &= ~(1 << OVS_KEY_ATTR_ICMP); + if (!fp) + return false; - *key_len = SW_FLOW_KEY_OFFSET(ipv4.tp); - icmp_key = nla_data(a[OVS_KEY_ATTR_ICMP]); - swkey->ipv4.tp.src = htons(icmp_key->icmp_type); - swkey->ipv4.tp.dst = htons(icmp_key->icmp_code); - break; - } + for (i = 0; i < size; i++) + if (fp[i]) + return false; - return 0; + return true; } -static int ipv6_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_len, - const struct nlattr *a[], u64 *attrs) +static int __parse_flow_nlattrs(const struct nlattr *attr, + const struct nlattr *a[], + u64 *attrsp, bool nz) { - const struct ovs_key_icmpv6 *icmpv6_key; - const struct ovs_key_tcp *tcp_key; - const struct ovs_key_udp *udp_key; + const struct nlattr *nla; + u64 attrs; + int rem; - switch (swkey->ip.proto) { - case IPPROTO_TCP: - if (!(*attrs & (1 << OVS_KEY_ATTR_TCP))) - return -EINVAL; - *attrs &= ~(1 << OVS_KEY_ATTR_TCP); + attrs = *attrsp; + nla_for_each_nested(nla, attr, rem) { + u16 type = nla_type(nla); + int expected_len; - *key_len = SW_FLOW_KEY_OFFSET(ipv6.tp); - tcp_key = nla_data(a[OVS_KEY_ATTR_TCP]); - swkey->ipv6.tp.src = tcp_key->tcp_src; - swkey->ipv6.tp.dst = tcp_key->tcp_dst; - break; + if (type > OVS_KEY_ATTR_MAX) { + OVS_NLERR("Unknown key attribute (type=%d, max=%d).\n", + type, OVS_KEY_ATTR_MAX); + } - case IPPROTO_UDP: - if (!(*attrs & (1 << OVS_KEY_ATTR_UDP))) + if (attrs & (1ULL << type)) { + OVS_NLERR("Duplicate key attribute (type %d).\n", type); return -EINVAL; - *attrs &= ~(1 << OVS_KEY_ATTR_UDP); - - *key_len = SW_FLOW_KEY_OFFSET(ipv6.tp); - udp_key = nla_data(a[OVS_KEY_ATTR_UDP]); - swkey->ipv6.tp.src = udp_key->udp_src; - swkey->ipv6.tp.dst = udp_key->udp_dst; - break; + } - case IPPROTO_ICMPV6: - if (!(*attrs & (1 << OVS_KEY_ATTR_ICMPV6))) + expected_len = ovs_key_lens[type]; + if (nla_len(nla) != expected_len && expected_len != -1) { + OVS_NLERR("Key attribute has unexpected length (type=%d" + ", length=%d, expected=%d).\n", type, + nla_len(nla), expected_len); return -EINVAL; - *attrs &= ~(1 << OVS_KEY_ATTR_ICMPV6); + } - *key_len = SW_FLOW_KEY_OFFSET(ipv6.tp); - icmpv6_key = nla_data(a[OVS_KEY_ATTR_ICMPV6]); - swkey->ipv6.tp.src = htons(icmpv6_key->icmpv6_type); - swkey->ipv6.tp.dst = htons(icmpv6_key->icmpv6_code); - - if (swkey->ipv6.tp.src == htons(NDISC_NEIGHBOUR_SOLICITATION) || - swkey->ipv6.tp.src == htons(NDISC_NEIGHBOUR_ADVERTISEMENT)) { - const struct ovs_key_nd *nd_key; - - if (!(*attrs & (1 << OVS_KEY_ATTR_ND))) - return -EINVAL; - *attrs &= ~(1 << OVS_KEY_ATTR_ND); - - *key_len = SW_FLOW_KEY_OFFSET(ipv6.nd); - nd_key = nla_data(a[OVS_KEY_ATTR_ND]); - memcpy(&swkey->ipv6.nd.target, nd_key->nd_target, - sizeof(swkey->ipv6.nd.target)); - memcpy(swkey->ipv6.nd.sll, nd_key->nd_sll, ETH_ALEN); - memcpy(swkey->ipv6.nd.tll, nd_key->nd_tll, ETH_ALEN); + if (!nz || !is_all_zero(nla_data(nla), expected_len)) { + attrs |= 1ULL << type; + a[type] = nla; } - break; + } + if (rem) { + OVS_NLERR("Message has %d unknown bytes.\n", rem); + return -EINVAL; } + *attrsp = attrs; return 0; } +static int parse_flow_mask_nlattrs(const struct nlattr *attr, + const struct nlattr *a[], u64 *attrsp) +{ + return __parse_flow_nlattrs(attr, a, attrsp, true); +} + static int parse_flow_nlattrs(const struct nlattr *attr, const struct nlattr *a[], u64 *attrsp) { - const struct nlattr *nla; - u64 attrs; - int rem; + return __parse_flow_nlattrs(attr, a, attrsp, false); +} - attrs = 0; - nla_for_each_nested(nla, attr, rem) { - u16 type = nla_type(nla); - int expected_len; +int ipv4_tun_from_nlattr(const struct nlattr *attr, + struct sw_flow_match *match, bool is_mask) +{ + struct nlattr *a; + int rem; + bool ttl = false; + __be16 tun_flags = 0; + + nla_for_each_nested(a, attr, rem) { + int type = nla_type(a); + static const u32 ovs_tunnel_key_lens[OVS_TUNNEL_KEY_ATTR_MAX + 1] = { + [OVS_TUNNEL_KEY_ATTR_ID] = sizeof(u64), + [OVS_TUNNEL_KEY_ATTR_IPV4_SRC] = sizeof(u32), + [OVS_TUNNEL_KEY_ATTR_IPV4_DST] = sizeof(u32), + [OVS_TUNNEL_KEY_ATTR_TOS] = 1, + [OVS_TUNNEL_KEY_ATTR_TTL] = 1, + [OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT] = 0, + [OVS_TUNNEL_KEY_ATTR_CSUM] = 0, + }; + + if (type > OVS_TUNNEL_KEY_ATTR_MAX) { + OVS_NLERR("Unknown IPv4 tunnel attribute (type=%d, max=%d).\n", + type, OVS_TUNNEL_KEY_ATTR_MAX); + return -EINVAL; + } - if (type > OVS_KEY_ATTR_MAX || attrs & (1ULL << type)) + if (ovs_tunnel_key_lens[type] != nla_len(a)) { + OVS_NLERR("IPv4 tunnel attribute type has unexpected " + " legnth (type=%d, length=%d, expected=%d).\n", + type, nla_len(a), ovs_tunnel_key_lens[type]); return -EINVAL; + } - expected_len = ovs_key_lens[type]; - if (nla_len(nla) != expected_len && expected_len != -1) + switch (type) { + case OVS_TUNNEL_KEY_ATTR_ID: + SW_FLOW_KEY_PUT(match, tun_key.tun_id, + nla_get_be64(a), is_mask); + tun_flags |= TUNNEL_KEY; + break; + case OVS_TUNNEL_KEY_ATTR_IPV4_SRC: + SW_FLOW_KEY_PUT(match, tun_key.ipv4_src, + nla_get_be32(a), is_mask); + break; + case OVS_TUNNEL_KEY_ATTR_IPV4_DST: + SW_FLOW_KEY_PUT(match, tun_key.ipv4_dst, + nla_get_be32(a), is_mask); + break; + case OVS_TUNNEL_KEY_ATTR_TOS: + SW_FLOW_KEY_PUT(match, tun_key.ipv4_tos, + nla_get_u8(a), is_mask); + break; + case OVS_TUNNEL_KEY_ATTR_TTL: + SW_FLOW_KEY_PUT(match, tun_key.ipv4_ttl, + nla_get_u8(a), is_mask); + ttl = true; + break; + case OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT: + tun_flags |= TUNNEL_DONT_FRAGMENT; + break; + case OVS_TUNNEL_KEY_ATTR_CSUM: + tun_flags |= TUNNEL_CSUM; + break; + default: return -EINVAL; + } + } - attrs |= 1ULL << type; - a[type] = nla; + SW_FLOW_KEY_PUT(match, tun_key.tun_flags, tun_flags, is_mask); + + if (rem > 0) { + OVS_NLERR("IPv4 tunnel attribute has %d unknown bytes.\n", rem); + return -EINVAL; } - if (rem) + + if (!match->key->tun_key.ipv4_dst) { + OVS_NLERR("IPv4 tunnel destination address is zero.\n"); return -EINVAL; + } + + if (!ttl) { + OVS_NLERR("IPv4 tunnel TTL not specified.\n"); + return -EINVAL; + } - *attrsp = attrs; 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) +int ipv4_tun_to_nlattr(struct sk_buff *skb, + const struct ovs_key_ipv4_tunnel *tun_key, + const struct ovs_key_ipv4_tunnel *output) { - const struct nlattr *a[OVS_KEY_ATTR_MAX + 1]; - const struct ovs_key_ethernet *eth_key; - int key_len; - u64 attrs; - int err; + struct nlattr *nla; - memset(swkey, 0, sizeof(struct sw_flow_key)); - key_len = SW_FLOW_KEY_OFFSET(eth); + nla = nla_nest_start(skb, OVS_KEY_ATTR_TUNNEL); + if (!nla) + return -EMSGSIZE; + + if (output->tun_flags & TUNNEL_KEY && + nla_put_be64(skb, OVS_TUNNEL_KEY_ATTR_ID, output->tun_id)) + return -EMSGSIZE; + if (output->ipv4_src && + nla_put_be32(skb, OVS_TUNNEL_KEY_ATTR_IPV4_SRC, output->ipv4_src)) + return -EMSGSIZE; + if (output->ipv4_dst && + nla_put_be32(skb, OVS_TUNNEL_KEY_ATTR_IPV4_DST, output->ipv4_dst)) + return -EMSGSIZE; + if (output->ipv4_tos && + nla_put_u8(skb, OVS_TUNNEL_KEY_ATTR_TOS, output->ipv4_tos)) + return -EMSGSIZE; + if (nla_put_u8(skb, OVS_TUNNEL_KEY_ATTR_TTL, output->ipv4_ttl)) + return -EMSGSIZE; + if ((output->tun_flags & TUNNEL_DONT_FRAGMENT) && + nla_put_flag(skb, OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT)) + return -EMSGSIZE; + if ((output->tun_flags & TUNNEL_CSUM) && + nla_put_flag(skb, OVS_TUNNEL_KEY_ATTR_CSUM)) + return -EMSGSIZE; + + nla_nest_end(skb, nla); + return 0; +} - 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); +static int metadata_from_nlattrs(struct sw_flow_match *match, u64 *attrs, + const struct nlattr **a, bool is_mask) +{ + if (*attrs & (1ULL << OVS_KEY_ATTR_PRIORITY)) { + SW_FLOW_KEY_PUT(match, phy.priority, + nla_get_u32(a[OVS_KEY_ATTR_PRIORITY]), is_mask); + *attrs &= ~(1ULL << OVS_KEY_ATTR_PRIORITY); } - if (attrs & (1 << OVS_KEY_ATTR_IN_PORT)) { + + if (*attrs & (1ULL << OVS_KEY_ATTR_IN_PORT)) { u32 in_port = nla_get_u32(a[OVS_KEY_ATTR_IN_PORT]); - if (in_port >= DP_MAX_PORTS) + + if (!is_mask && 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 = USHRT_MAX; + SW_FLOW_KEY_PUT(match, phy.in_port, in_port, is_mask); + *attrs &= ~(1ULL << OVS_KEY_ATTR_IN_PORT); + } else if (!is_mask) { + SW_FLOW_KEY_PUT(match, phy.in_port, DP_MAX_PORTS, is_mask); } - if (attrs & (1ULL << OVS_KEY_ATTR_TUN_ID)) { - swkey->phy.tun_id = nla_get_be64(a[OVS_KEY_ATTR_TUN_ID]); - attrs &= ~(1ULL << OVS_KEY_ATTR_TUN_ID); + if (*attrs & (1ULL << OVS_KEY_ATTR_SKB_MARK)) { + uint32_t mark = nla_get_u32(a[OVS_KEY_ATTR_SKB_MARK]); +#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,20) && !defined(CONFIG_NETFILTER) + if (!is_mask && mark != 0) { + OVS_NLERR("skb->mark must be zero on this kernel (mark=%d).\n", mark); + return -EINVAL; + } +#endif + SW_FLOW_KEY_PUT(match, phy.skb_mark, mark, is_mask); + *attrs &= ~(1ULL << OVS_KEY_ATTR_SKB_MARK); + } + if (*attrs & (1ULL << OVS_KEY_ATTR_TUNNEL)) { + if (ipv4_tun_from_nlattr(a[OVS_KEY_ATTR_TUNNEL], match, + is_mask)) + return -EINVAL; + *attrs &= ~(1ULL << OVS_KEY_ATTR_TUNNEL); } + return 0; +} - /* Data attributes. */ - if (!(attrs & (1 << OVS_KEY_ATTR_ETHERNET))) - return -EINVAL; - attrs &= ~(1 << OVS_KEY_ATTR_ETHERNET); +static int ovs_key_from_nlattrs(struct sw_flow_match *match, u64 attrs, + const struct nlattr **a, bool is_mask) +{ + int err; + u64 orig_attrs = attrs; + + err = metadata_from_nlattrs(match, &attrs, a, is_mask); + if (err) + return err; - 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 & (1ULL << OVS_KEY_ATTR_ETHERNET)) { + const struct ovs_key_ethernet *eth_key; + + eth_key = nla_data(a[OVS_KEY_ATTR_ETHERNET]); + SW_FLOW_KEY_MEMCPY(match, eth.src, + eth_key->eth_src, ETH_ALEN, is_mask); + SW_FLOW_KEY_MEMCPY(match, eth.dst, + eth_key->eth_dst, ETH_ALEN, is_mask); + attrs &= ~(1ULL << OVS_KEY_ATTR_ETHERNET); + } - if (attrs & (1u << OVS_KEY_ATTR_ETHERTYPE) && - nla_get_be16(a[OVS_KEY_ATTR_ETHERTYPE]) == htons(ETH_P_8021Q)) { - const struct nlattr *encap; + if (attrs & (1ULL << OVS_KEY_ATTR_VLAN)) { __be16 tci; - if (attrs != ((1 << OVS_KEY_ATTR_VLAN) | - (1 << OVS_KEY_ATTR_ETHERTYPE) | - (1 << OVS_KEY_ATTR_ENCAP))) + tci = nla_get_be16(a[OVS_KEY_ATTR_VLAN]); + if (!(tci & htons(VLAN_TAG_PRESENT))) { + if (is_mask) + OVS_NLERR("VLAN TCI mask does not have exact match for VLAN_TAG_PRESENT bit.\n"); + else + OVS_NLERR("VLAN TCI does not have VLAN_TAG_PRESENT bit set.\n"); + 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; + SW_FLOW_KEY_PUT(match, eth.tci, tci, is_mask); + attrs &= ~(1ULL << OVS_KEY_ATTR_VLAN); + } else if (!is_mask) + SW_FLOW_KEY_PUT(match, eth.tci, htons(0xffff), true); - 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; + if (attrs & (1ULL << OVS_KEY_ATTR_ETHERTYPE)) { + __be16 eth_type; - swkey->eth.type = htons(ETH_P_8021Q); - *key_lenp = key_len; - return 0; - } else { + eth_type = nla_get_be16(a[OVS_KEY_ATTR_ETHERTYPE]); + if (!is_mask && ntohs(eth_type) < ETH_P_802_3_MIN) { + OVS_NLERR("EtherType is less than mimimum (type=%x, min=%x).\n", + ntohs(eth_type), ETH_P_802_3_MIN); 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) < 1536) - return -EINVAL; - attrs &= ~(1 << OVS_KEY_ATTR_ETHERTYPE); - } else { - swkey->eth.type = htons(ETH_P_802_2); + SW_FLOW_KEY_PUT(match, eth.type, eth_type, is_mask); + attrs &= ~(1ULL << OVS_KEY_ATTR_ETHERTYPE); + } else if (!is_mask) { + SW_FLOW_KEY_PUT(match, eth.type, htons(ETH_P_802_2), is_mask); } - if (swkey->eth.type == htons(ETH_P_IP)) { + if (attrs & (1ULL << OVS_KEY_ATTR_IPV4)) { 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) + if (!is_mask && ipv4_key->ipv4_frag > OVS_FRAG_TYPE_MAX) { + OVS_NLERR("Unknown IPv4 fragment type (value=%d, max=%d).\n", + 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)) { + SW_FLOW_KEY_PUT(match, ip.proto, + ipv4_key->ipv4_proto, is_mask); + SW_FLOW_KEY_PUT(match, ip.tos, + ipv4_key->ipv4_tos, is_mask); + SW_FLOW_KEY_PUT(match, ip.ttl, + ipv4_key->ipv4_ttl, is_mask); + SW_FLOW_KEY_PUT(match, ip.frag, + ipv4_key->ipv4_frag, is_mask); + SW_FLOW_KEY_PUT(match, ipv4.addr.src, + ipv4_key->ipv4_src, is_mask); + SW_FLOW_KEY_PUT(match, ipv4.addr.dst, + ipv4_key->ipv4_dst, is_mask); + attrs &= ~(1ULL << OVS_KEY_ATTR_IPV4); + } + + if (attrs & (1ULL << OVS_KEY_ATTR_IPV6)) { const struct ovs_key_ipv6 *ipv6_key; - if (!(attrs & (1 << OVS_KEY_ATTR_IPV6))) + ipv6_key = nla_data(a[OVS_KEY_ATTR_IPV6]); + if (!is_mask && ipv6_key->ipv6_frag > OVS_FRAG_TYPE_MAX) { + OVS_NLERR("Unknown IPv6 fragment type (value=%d, max=%d).\n", + ipv6_key->ipv6_frag, OVS_FRAG_TYPE_MAX); return -EINVAL; - attrs &= ~(1 << OVS_KEY_ATTR_IPV6); + } + SW_FLOW_KEY_PUT(match, ipv6.label, + ipv6_key->ipv6_label, is_mask); + SW_FLOW_KEY_PUT(match, ip.proto, + ipv6_key->ipv6_proto, is_mask); + SW_FLOW_KEY_PUT(match, ip.tos, + ipv6_key->ipv6_tclass, is_mask); + SW_FLOW_KEY_PUT(match, ip.ttl, + ipv6_key->ipv6_hlimit, is_mask); + SW_FLOW_KEY_PUT(match, ip.frag, + ipv6_key->ipv6_frag, is_mask); + SW_FLOW_KEY_MEMCPY(match, ipv6.addr.src, + ipv6_key->ipv6_src, + sizeof(match->key->ipv6.addr.src), + is_mask); + SW_FLOW_KEY_MEMCPY(match, ipv6.addr.dst, + ipv6_key->ipv6_dst, + sizeof(match->key->ipv6.addr.dst), + is_mask); + + attrs &= ~(1ULL << 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) + if (attrs & (1ULL << OVS_KEY_ATTR_ARP)) { + const struct ovs_key_arp *arp_key; + + arp_key = nla_data(a[OVS_KEY_ATTR_ARP]); + if (!is_mask && (arp_key->arp_op & htons(0xff00))) { + OVS_NLERR("Unknown ARP opcode (opcode=%d).\n", + arp_key->arp_op); 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); + } + + SW_FLOW_KEY_PUT(match, ipv4.addr.src, + arp_key->arp_sip, is_mask); + SW_FLOW_KEY_PUT(match, ipv4.addr.dst, + arp_key->arp_tip, is_mask); + SW_FLOW_KEY_PUT(match, ip.proto, + ntohs(arp_key->arp_op), is_mask); + SW_FLOW_KEY_MEMCPY(match, ipv4.arp.sha, + arp_key->arp_sha, ETH_ALEN, is_mask); + SW_FLOW_KEY_MEMCPY(match, ipv4.arp.tha, + arp_key->arp_tha, ETH_ALEN, is_mask); + + attrs &= ~(1ULL << OVS_KEY_ATTR_ARP); + } + + if (attrs & (1ULL << OVS_KEY_ATTR_TCP)) { + const struct ovs_key_tcp *tcp_key; + + tcp_key = nla_data(a[OVS_KEY_ATTR_TCP]); + if (orig_attrs & (1ULL << OVS_KEY_ATTR_IPV4)) { + SW_FLOW_KEY_PUT(match, ipv4.tp.src, + tcp_key->tcp_src, is_mask); + SW_FLOW_KEY_PUT(match, ipv4.tp.dst, + tcp_key->tcp_dst, is_mask); + } else { + SW_FLOW_KEY_PUT(match, ipv6.tp.src, + tcp_key->tcp_src, is_mask); + SW_FLOW_KEY_PUT(match, ipv6.tp.dst, + tcp_key->tcp_dst, is_mask); + } + attrs &= ~(1ULL << OVS_KEY_ATTR_TCP); + } + + if (attrs & (1ULL << OVS_KEY_ATTR_UDP)) { + const struct ovs_key_udp *udp_key; + + udp_key = nla_data(a[OVS_KEY_ATTR_UDP]); + if (orig_attrs & (1ULL << OVS_KEY_ATTR_IPV4)) { + SW_FLOW_KEY_PUT(match, ipv4.tp.src, + udp_key->udp_src, is_mask); + SW_FLOW_KEY_PUT(match, ipv4.tp.dst, + udp_key->udp_dst, is_mask); + } else { + SW_FLOW_KEY_PUT(match, ipv6.tp.src, + udp_key->udp_src, is_mask); + SW_FLOW_KEY_PUT(match, ipv6.tp.dst, + udp_key->udp_dst, is_mask); + } + attrs &= ~(1ULL << OVS_KEY_ATTR_UDP); + } + + if (attrs & (1ULL << OVS_KEY_ATTR_ICMP)) { + const struct ovs_key_icmp *icmp_key; + + icmp_key = nla_data(a[OVS_KEY_ATTR_ICMP]); + SW_FLOW_KEY_PUT(match, ipv4.tp.src, + htons(icmp_key->icmp_type), is_mask); + SW_FLOW_KEY_PUT(match, ipv4.tp.dst, + htons(icmp_key->icmp_code), is_mask); + attrs &= ~(1ULL << OVS_KEY_ATTR_ICMP); + } + + if (attrs & (1ULL << OVS_KEY_ATTR_ICMPV6)) { + const struct ovs_key_icmpv6 *icmpv6_key; + + icmpv6_key = nla_data(a[OVS_KEY_ATTR_ICMPV6]); + SW_FLOW_KEY_PUT(match, ipv6.tp.src, + htons(icmpv6_key->icmpv6_type), is_mask); + SW_FLOW_KEY_PUT(match, ipv6.tp.dst, + htons(icmpv6_key->icmpv6_code), is_mask); + attrs &= ~(1ULL << OVS_KEY_ATTR_ICMPV6); + } + + if (attrs & (1ULL << OVS_KEY_ATTR_ND)) { + const struct ovs_key_nd *nd_key; + + nd_key = nla_data(a[OVS_KEY_ATTR_ND]); + SW_FLOW_KEY_MEMCPY(match, ipv6.nd.target, + nd_key->nd_target, + sizeof(match->key->ipv6.nd.target), + is_mask); + SW_FLOW_KEY_MEMCPY(match, ipv6.nd.sll, + nd_key->nd_sll, ETH_ALEN, is_mask); + SW_FLOW_KEY_MEMCPY(match, ipv6.nd.tll, + nd_key->nd_tll, ETH_ALEN, is_mask); + attrs &= ~(1ULL << OVS_KEY_ATTR_ND); + } + + if (attrs != 0) + return -EINVAL; + + return 0; +} + +/** + * ovs_match_from_nlattrs - parses Netlink attributes into a flow key and + * mask. In case the 'mask' is NULL, the flow is treated as exact match + * flow. Otherwise, it is treated as a wildcarded flow, except the mask + * does not include any don't care bit. + * @match: receives the extracted flow match information. + * @key: Netlink attribute holding nested %OVS_KEY_ATTR_* Netlink attribute + * sequence. The fields should of the packet that triggered the creation + * of this flow. + * @mask: Optional. Netlink attribute holding nested %OVS_KEY_ATTR_* Netlink + * attribute specifies the mask field of the wildcarded flow. + */ +int ovs_match_from_nlattrs(struct sw_flow_match *match, + const struct nlattr *key, + const struct nlattr *mask) +{ + const struct nlattr *a[OVS_KEY_ATTR_MAX + 1]; + const struct nlattr *encap; + u64 key_attrs = 0; + u64 mask_attrs = 0; + bool encap_valid = false; + int err; + + err = parse_flow_nlattrs(key, a, &key_attrs); + if (err) + return err; + + if (key_attrs & 1ULL << OVS_KEY_ATTR_ENCAP) { + encap = a[OVS_KEY_ATTR_ENCAP]; + key_attrs &= ~(1ULL << OVS_KEY_ATTR_ENCAP); + if (nla_len(encap)) { + __be16 eth_type = 0; /* ETH_P_8021Q */ + + if (a[OVS_KEY_ATTR_ETHERTYPE]) + eth_type = nla_get_be16(a[OVS_KEY_ATTR_ETHERTYPE]); + + if ((eth_type == htons(ETH_P_8021Q)) && (a[OVS_KEY_ATTR_VLAN])) { + encap_valid = true; + key_attrs &= ~(1ULL << OVS_KEY_ATTR_ETHERTYPE); + err = parse_flow_nlattrs(encap, a, &key_attrs); + } else { + OVS_NLERR("Encap attribute is set for a non-VLAN frame.\n"); + err = -EINVAL; + } + if (err) return err; } - } else if (swkey->eth.type == htons(ETH_P_ARP)) { - const struct ovs_key_arp *arp_key; + } - if (!(attrs & (1 << OVS_KEY_ATTR_ARP))) - return -EINVAL; - attrs &= ~(1 << OVS_KEY_ATTR_ARP); + err = ovs_key_from_nlattrs(match, key_attrs, a, false); + if (err) + return err; - 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 (mask) { + err = parse_flow_mask_nlattrs(mask, a, &mask_attrs); + if (err) + return err; + + if ((mask_attrs & 1ULL << OVS_KEY_ATTR_ENCAP) && encap_valid) { + __be16 eth_type = 0; + + mask_attrs &= ~(1ULL << OVS_KEY_ATTR_ENCAP); + if (a[OVS_KEY_ATTR_ETHERTYPE]) + eth_type = nla_get_be16(a[OVS_KEY_ATTR_ETHERTYPE]); + if (eth_type == htons(0xffff)) { + mask_attrs &= ~(1ULL << OVS_KEY_ATTR_ETHERTYPE); + encap = a[OVS_KEY_ATTR_ENCAP]; + err = parse_flow_mask_nlattrs(encap, a, &mask_attrs); + } else { + OVS_NLERR("VLAN frames must have an exact match" + " on the TPID (mask=%x).\n", + ntohs(eth_type)); + err = -EINVAL; + } + + if (err) + return err; + } + + err = ovs_key_from_nlattrs(match, mask_attrs, a, true); + if (err) + return err; + } else { + /* Populate exact match flow's key mask. */ + if (match->mask) + ovs_sw_flow_mask_set(match->mask, &match->range, 0xff); } - if (attrs) + if (!ovs_match_validate(match, key_attrs, mask_attrs)) return -EINVAL; - *key_lenp = key_len; return 0; } /** * ovs_flow_metadata_from_nlattrs - parses Netlink attributes into a flow key. - * @in_port: receives the extracted input port. - * @tun_id: receives the extracted tunnel ID. - * @key: Netlink attribute holding nested %OVS_KEY_ATTR_* Netlink attribute + * @flow: Receives extracted in_port, priority, tun_key and skb_mark. + * @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 @@ -1164,80 +1662,100 @@ int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp, * 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(u32 *priority, u16 *in_port, __be64 *tun_id, - const struct nlattr *attr) + +int ovs_flow_metadata_from_nlattrs(struct sw_flow *flow, + const struct nlattr *attr) { - const struct nlattr *nla; - int rem; + struct ovs_key_ipv4_tunnel *tun_key = &flow->key.tun_key; + const struct nlattr *a[OVS_KEY_ATTR_MAX + 1]; + u64 attrs = 0; + int err; + struct sw_flow_match match; - *in_port = USHRT_MAX; - *tun_id = 0; - *priority = 0; + 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) { - if (nla_len(nla) != ovs_key_lens[type]) - return -EINVAL; - - switch (type) { - case OVS_KEY_ATTR_PRIORITY: - *priority = nla_get_u32(nla); - break; - - case OVS_KEY_ATTR_TUN_ID: - *tun_id = nla_get_be64(nla); - break; - - case OVS_KEY_ATTR_IN_PORT: - if (nla_get_u32(nla) >= DP_MAX_PORTS) - return -EINVAL; - *in_port = nla_get_u32(nla); - break; - } - } - } - if (rem) + err = parse_flow_nlattrs(attr, a, &attrs); + if (err) return -EINVAL; + + memset(&match, 0, sizeof(match)); + match.key = &flow->key; + + err = metadata_from_nlattrs(&match, &attrs, a, false); + if (err) + return err; + return 0; } -int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) +int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, + const struct sw_flow_key *output, 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); + if (nla_put_u32(skb, OVS_KEY_ATTR_PRIORITY, output->phy.priority)) + goto nla_put_failure; - if (swkey->phy.tun_id != cpu_to_be64(0)) - NLA_PUT_BE64(skb, OVS_KEY_ATTR_TUN_ID, swkey->phy.tun_id); + if (swkey->tun_key.ipv4_dst && + ipv4_tun_to_nlattr(skb, &swkey->tun_key, &output->tun_key)) + goto nla_put_failure; - if (swkey->phy.in_port != USHRT_MAX) - NLA_PUT_U32(skb, OVS_KEY_ATTR_IN_PORT, swkey->phy.in_port); + if (swkey->phy.in_port == DP_MAX_PORTS) { + if ((swkey != output) && (output->phy.in_port == 0xffff)) + if (nla_put_u32(skb, OVS_KEY_ATTR_IN_PORT, 0xffffffff)) + goto nla_put_failure; + } else { + u16 upper_u16; + upper_u16 = (swkey == output) ? 0 : 0xffff; + + if (nla_put_u32(skb, OVS_KEY_ATTR_IN_PORT, + (upper_u16 << 16) | output->phy.in_port)) + goto nla_put_failure; + } + + if (nla_put_u32(skb, OVS_KEY_ATTR_SKB_MARK, output->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); + memcpy(eth_key->eth_src, output->eth.src, ETH_ALEN); + memcpy(eth_key->eth_dst, output->eth.dst, ETH_ALEN); if (swkey->eth.tci || swkey->eth.type == htons(ETH_P_8021Q)) { - NLA_PUT_BE16(skb, OVS_KEY_ATTR_ETHERTYPE, htons(ETH_P_8021Q)); - NLA_PUT_BE16(skb, OVS_KEY_ATTR_VLAN, swkey->eth.tci); + __be16 eth_type; + eth_type = (swkey == output) ? htons(ETH_P_8021Q) : htons(0xffff) ; + if (nla_put_be16(skb, OVS_KEY_ATTR_ETHERTYPE, eth_type) || + nla_put_be16(skb, OVS_KEY_ATTR_VLAN, output->eth.tci)) + goto nla_put_failure; encap = nla_nest_start(skb, OVS_KEY_ATTR_ENCAP); if (!swkey->eth.tci) goto unencap; - } else { + } else encap = NULL; - } - if (swkey->eth.type == htons(ETH_P_802_2)) + if (swkey->eth.type == htons(ETH_P_802_2)) { + /* + * Ethertype 802.2 is represented in the netlink with omitted + * OVS_KEY_ATTR_ETHERTYPE in the flow key attribute, and + * 0xffff in the mask attribute. Ethertype can also + * be wildcarded. + */ + if (swkey != output && output->eth.type) + if (nla_put_be16(skb, OVS_KEY_ATTR_ETHERTYPE, + output->eth.type)) + goto nla_put_failure; goto unencap; + } - NLA_PUT_BE16(skb, OVS_KEY_ATTR_ETHERTYPE, swkey->eth.type); + if (nla_put_be16(skb, OVS_KEY_ATTR_ETHERTYPE, output->eth.type)) + goto nla_put_failure; if (swkey->eth.type == htons(ETH_P_IP)) { struct ovs_key_ipv4 *ipv4_key; @@ -1246,12 +1764,12 @@ int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) 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; + ipv4_key->ipv4_src = output->ipv4.addr.src; + ipv4_key->ipv4_dst = output->ipv4.addr.dst; + ipv4_key->ipv4_proto = output->ip.proto; + ipv4_key->ipv4_tos = output->ip.tos; + ipv4_key->ipv4_ttl = output->ip.ttl; + ipv4_key->ipv4_frag = output->ip.frag; } else if (swkey->eth.type == htons(ETH_P_IPV6)) { struct ovs_key_ipv6 *ipv6_key; @@ -1259,16 +1777,17 @@ int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) if (!nla) goto nla_put_failure; ipv6_key = nla_data(nla); - memcpy(ipv6_key->ipv6_src, &swkey->ipv6.addr.src, + memcpy(ipv6_key->ipv6_src, &output->ipv6.addr.src, sizeof(ipv6_key->ipv6_src)); - memcpy(ipv6_key->ipv6_dst, &swkey->ipv6.addr.dst, + memcpy(ipv6_key->ipv6_dst, &output->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)) { + ipv6_key->ipv6_label = output->ipv6.label; + ipv6_key->ipv6_proto = output->ip.proto; + ipv6_key->ipv6_tclass = output->ip.tos; + ipv6_key->ipv6_hlimit = output->ip.ttl; + ipv6_key->ipv6_frag = output->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)); @@ -1276,11 +1795,11 @@ int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) 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); + arp_key->arp_sip = output->ipv4.addr.src; + arp_key->arp_tip = output->ipv4.addr.dst; + arp_key->arp_op = htons(output->ip.proto); + memcpy(arp_key->arp_sha, output->ipv4.arp.sha, ETH_ALEN); + memcpy(arp_key->arp_tha, output->ipv4.arp.tha, ETH_ALEN); } if ((swkey->eth.type == htons(ETH_P_IP) || @@ -1295,11 +1814,11 @@ int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) 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; + tcp_key->tcp_src = output->ipv4.tp.src; + tcp_key->tcp_dst = output->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; + tcp_key->tcp_src = output->ipv6.tp.src; + tcp_key->tcp_dst = output->ipv6.tp.dst; } } else if (swkey->ip.proto == IPPROTO_UDP) { struct ovs_key_udp *udp_key; @@ -1309,11 +1828,11 @@ int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) 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; + udp_key->udp_src = output->ipv4.tp.src; + udp_key->udp_dst = output->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; + udp_key->udp_src = output->ipv6.tp.src; + udp_key->udp_dst = output->ipv6.tp.dst; } } else if (swkey->eth.type == htons(ETH_P_IP) && swkey->ip.proto == IPPROTO_ICMP) { @@ -1323,8 +1842,8 @@ int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) 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); + icmp_key->icmp_type = ntohs(output->ipv4.tp.src); + icmp_key->icmp_code = ntohs(output->ipv4.tp.dst); } else if (swkey->eth.type == htons(ETH_P_IPV6) && swkey->ip.proto == IPPROTO_ICMPV6) { struct ovs_key_icmpv6 *icmpv6_key; @@ -1334,8 +1853,8 @@ int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) 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); + icmpv6_key->icmpv6_type = ntohs(output->ipv6.tp.src); + icmpv6_key->icmpv6_code = ntohs(output->ipv6.tp.dst); if (icmpv6_key->icmpv6_type == NDISC_NEIGHBOUR_SOLICITATION || icmpv6_key->icmpv6_type == NDISC_NEIGHBOUR_ADVERTISEMENT) { @@ -1345,10 +1864,10 @@ int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) if (!nla) goto nla_put_failure; nd_key = nla_data(nla); - memcpy(nd_key->nd_target, &swkey->ipv6.nd.target, + memcpy(nd_key->nd_target, &output->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); + memcpy(nd_key->nd_sll, output->ipv6.nd.sll, ETH_ALEN); + memcpy(nd_key->nd_tll, output->ipv6.nd.tll, ETH_ALEN); } } } @@ -1380,3 +1899,91 @@ void ovs_flow_exit(void) { kmem_cache_destroy(flow_cache); } + +struct sw_flow_mask *ovs_sw_flow_mask_alloc(void) +{ + struct sw_flow_mask *mask; + + mask = kmalloc(sizeof(*mask), GFP_KERNEL); + if (mask) + mask->ref_count = 0; + + return mask; +} + +void ovs_sw_flow_mask_add_ref(struct sw_flow_mask *mask) +{ + mask->ref_count++; +} + +static void rcu_free_sw_flow_mask_cb(struct rcu_head *rcu) +{ + struct sw_flow_mask *mask = container_of(rcu, struct sw_flow_mask, rcu); + + kfree(mask); +} + +void ovs_sw_flow_mask_del_ref(struct sw_flow_mask *mask, bool deferred) +{ + if (!mask) + return; + + BUG_ON(!mask->ref_count); + mask->ref_count--; + + if (!mask->ref_count) { + list_del_rcu(&mask->list); + if (deferred) + call_rcu(&mask->rcu, rcu_free_sw_flow_mask_cb); + else + kfree(mask); + } +} + +static bool ovs_sw_flow_mask_equal(const struct sw_flow_mask *a, + const struct sw_flow_mask *b) +{ + u8 *a_ = (u8 *)&a->key + a->range.start; + u8 *b_ = (u8 *)&b->key + b->range.start; + + return (a->range.end == b->range.end) + && (a->range.start == b->range.start) + && (memcmp(a_, b_, ovs_sw_flow_mask_actual_size(a)) == 0); +} + +struct sw_flow_mask *ovs_sw_flow_mask_find(const struct flow_table *tbl, + const struct sw_flow_mask *mask) +{ + struct list_head *ml; + + list_for_each(ml, tbl->mask_list) { + struct sw_flow_mask *m; + m = container_of(ml, struct sw_flow_mask, list); + if (ovs_sw_flow_mask_equal(mask, m)) + return m; + } + + return NULL; +} + +/** + * add a new mask into the mask list. + * The caller needs to make sure that 'mask' is not the same + * as any masks that are already on the list. + */ +void ovs_sw_flow_mask_insert(struct flow_table *tbl, struct sw_flow_mask *mask) +{ + list_add_rcu(&mask->list, tbl->mask_list); +} + +/** + * Set 'range' fields in the mask to the value of 'val'. + */ +static void ovs_sw_flow_mask_set(struct sw_flow_mask *mask, + struct sw_flow_key_range *range, u8 val) +{ + u8 *m = (u8 *)&mask->key + range->start; + + mask->range = *range; + memset(m, val, ovs_sw_flow_mask_size_roundup(mask)); +}