X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=datapath%2Fflow.c;h=c6f591afe166ecaf17ef98ca742b8bed991435e1;hb=6a542738b2a59f98831fa36e6208d388c007d0b8;hp=b0252b83f35edd616a751c56549f58ecdf7e6344;hpb=34118caeded9c72aa259ab76b97148efda7bdc22;p=openvswitch diff --git a/datapath/flow.c b/datapath/flow.c index b0252b83..c6f591af 100644 --- a/datapath/flow.c +++ b/datapath/flow.c @@ -1,9 +1,19 @@ /* - * Distributed under the terms of the GNU GPL version 2. - * Copyright (c) 2007, 2008, 2009, 2010, 2011 Nicira Networks. + * Copyright (c) 2007-2011 Nicira Networks. * - * Significant portions of this file may be copied from parts of the Linux - * kernel, by Linus Torvalds and others. + * This program is free software; you can redistribute it and/or + * modify it under the terms of version 2 of the GNU General Public + * License as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA + * 02110-1301, USA */ #include "flow.h" @@ -101,7 +111,7 @@ static bool icmphdr_ok(struct sk_buff *skb) sizeof(struct icmphdr)); } -u64 flow_used_time(unsigned long flow_jiffies) +u64 ovs_flow_used_time(unsigned long flow_jiffies) { struct timespec cur_ts; u64 cur_ms, idle_ms; @@ -202,8 +212,8 @@ static int parse_ipv6hdr(struct sk_buff *skb, struct sw_flow_key *key, key->ip.tos = ipv6_get_dsfield(nh); key->ip.ttl = nh->hop_limit; key->ipv6.label = *(__be32 *)nh & htonl(IPV6_FLOWINFO_FLOWLABEL); - ipv6_addr_copy(&key->ipv6.addr.src, &nh->saddr); - ipv6_addr_copy(&key->ipv6.addr.dst, &nh->daddr); + key->ipv6.addr.src = nh->saddr; + key->ipv6.addr.dst = nh->daddr; payload_ofs = skip_exthdr(skb, payload_ofs, &nexthdr, &key->ip.frag); if (unlikely(payload_ofs < 0)) @@ -224,7 +234,7 @@ static bool icmp6hdr_ok(struct sk_buff *skb) #define TCP_FLAGS_OFFSET 13 #define TCP_FLAG_MASK 0x3f -void flow_used(struct sw_flow *flow, struct sk_buff *skb) +void ovs_flow_used(struct sw_flow *flow, struct sk_buff *skb) { u8 tcp_flags = 0; @@ -242,7 +252,7 @@ void flow_used(struct sw_flow *flow, struct sk_buff *skb) spin_unlock(&flow->lock); } -struct sw_flow_actions *flow_actions_alloc(const struct nlattr *actions) +struct sw_flow_actions *ovs_flow_actions_alloc(const struct nlattr *actions) { int actions_len = nla_len(actions); struct sw_flow_actions *sfa; @@ -262,7 +272,7 @@ struct sw_flow_actions *flow_actions_alloc(const struct nlattr *actions) return sfa; } -struct sw_flow *flow_alloc(void) +struct sw_flow *ovs_flow_alloc(void) { struct sw_flow *flow; @@ -278,15 +288,15 @@ struct sw_flow *flow_alloc(void) return flow; } -static struct hlist_head __rcu *find_bucket(struct flow_table * table, u32 hash) +static struct hlist_head *find_bucket(struct flow_table *table, u32 hash) { return flex_array_get(table->buckets, (hash & (table->n_buckets - 1))); } -static struct flex_array __rcu *alloc_buckets(unsigned int n_buckets) +static struct flex_array *alloc_buckets(unsigned int n_buckets) { - struct flex_array __rcu *buckets; + struct flex_array *buckets; int i, err; buckets = flex_array_alloc(sizeof(struct hlist_head *), @@ -312,7 +322,7 @@ static void free_buckets(struct flex_array *buckets) flex_array_free(buckets); } -struct flow_table *flow_tbl_alloc(int new_size) +struct flow_table *ovs_flow_tbl_alloc(int new_size) { struct flow_table *table = kmalloc(sizeof(*table), GFP_KERNEL); @@ -334,10 +344,10 @@ struct flow_table *flow_tbl_alloc(int new_size) static void flow_free(struct sw_flow *flow) { flow->dead = true; - flow_put(flow); + ovs_flow_put(flow); } -void flow_tbl_destroy(struct flow_table *table) +void ovs_flow_tbl_destroy(struct flow_table *table) { int i; @@ -363,10 +373,10 @@ static void flow_tbl_destroy_rcu_cb(struct rcu_head *rcu) { struct flow_table *table = container_of(rcu, struct flow_table, rcu); - flow_tbl_destroy(table); + ovs_flow_tbl_destroy(table); } -void flow_tbl_deferred_destroy(struct flow_table *table) +void ovs_flow_tbl_deferred_destroy(struct flow_table *table) { if (!table) return; @@ -374,7 +384,7 @@ void flow_tbl_deferred_destroy(struct flow_table *table) call_rcu(&table->rcu, flow_tbl_destroy_rcu_cb); } -struct sw_flow *flow_tbl_next(struct flow_table *table, u32 *bucket, u32 *last) +struct sw_flow *ovs_flow_tbl_next(struct flow_table *table, u32 *bucket, u32 *last) { struct sw_flow *flow; struct hlist_head *head; @@ -399,13 +409,13 @@ struct sw_flow *flow_tbl_next(struct flow_table *table, u32 *bucket, u32 *last) return NULL; } -struct flow_table *flow_tbl_expand(struct flow_table *table) +struct flow_table *ovs_flow_tbl_expand(struct flow_table *table) { struct flow_table *new_table; int n_buckets = table->n_buckets * 2; int i; - new_table = flow_tbl_alloc(n_buckets); + new_table = ovs_flow_tbl_alloc(n_buckets); if (!new_table) return ERR_PTR(-ENOMEM); @@ -418,35 +428,35 @@ struct flow_table *flow_tbl_expand(struct flow_table *table) hlist_for_each_entry_safe(flow, n, pos, head, hash_node) { hlist_del_init_rcu(&flow->hash_node); - flow_tbl_insert(new_table, flow); + ovs_flow_tbl_insert(new_table, flow); } } return new_table; } -/* RCU callback used by flow_deferred_free. */ +/* RCU callback used by ovs_flow_deferred_free. */ static void rcu_free_flow_callback(struct rcu_head *rcu) { struct sw_flow *flow = container_of(rcu, struct sw_flow, rcu); flow->dead = true; - flow_put(flow); + ovs_flow_put(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 flow_deferred_free(struct sw_flow *flow) +void ovs_flow_deferred_free(struct sw_flow *flow) { call_rcu(&flow->rcu, rcu_free_flow_callback); } -void flow_hold(struct sw_flow *flow) +void ovs_flow_hold(struct sw_flow *flow) { atomic_inc(&flow->refcnt); } -void flow_put(struct sw_flow *flow) +void ovs_flow_put(struct sw_flow *flow) { if (unlikely(!flow)) return; @@ -457,7 +467,7 @@ void flow_put(struct sw_flow *flow) } } -/* RCU callback used by flow_deferred_free_acts. */ +/* RCU callback used by ovs_flow_deferred_free_acts. */ static void rcu_free_acts_callback(struct rcu_head *rcu) { struct sw_flow_actions *sf_acts = container_of(rcu, @@ -467,7 +477,7 @@ static void rcu_free_acts_callback(struct rcu_head *rcu) /* Schedules 'sf_acts' to be freed after the next RCU grace period. * The caller must hold rcu_read_lock for this to be sensible. */ -void flow_deferred_free_acts(struct sw_flow_actions *sf_acts) +void ovs_flow_deferred_free_acts(struct sw_flow_actions *sf_acts) { call_rcu(&sf_acts->rcu, rcu_free_acts_callback); } @@ -480,6 +490,9 @@ static int parse_vlan(struct sk_buff *skb, struct sw_flow_key *key) }; struct qtag_prefix *qp; + if (unlikely(skb->len < sizeof(struct qtag_prefix) + sizeof(__be16))) + return 0; + if (unlikely(!pskb_may_pull(skb, sizeof(struct qtag_prefix) + sizeof(__be16)))) return -ENOMEM; @@ -559,7 +572,7 @@ static int parse_icmpv6(struct sk_buff *skb, struct sw_flow_key *key, } nd = (struct nd_msg *)skb_transport_header(skb); - ipv6_addr_copy(&key->ipv6.nd.target, &nd->target); + key->ipv6.nd.target = nd->target; key_len = SW_FLOW_KEY_OFFSET(ipv6.nd); icmp_len -= sizeof(*nd); @@ -608,7 +621,7 @@ out: } /** - * flow_extract - extracts a flow key from an Ethernet frame. + * ovs_flow_extract - extracts a flow key from an Ethernet frame. * @skb: sk_buff that contains the frame, with skb->data pointing to the * Ethernet header * @in_port: port number on which @skb was received. @@ -631,7 +644,7 @@ out: * of a correct length, otherwise the same as skb->network_header. * For other key->dl_type values it is left untouched. */ -int flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key, +int ovs_flow_extract(struct sk_buff *skb, u16 in_port, struct sw_flow_key *key, int *key_lenp) { int error = 0; @@ -798,12 +811,12 @@ out: return error; } -u32 flow_hash(const struct sw_flow_key *key, int key_len) +u32 ovs_flow_hash(const struct sw_flow_key *key, int key_len) { return jhash2((u32 *)key, DIV_ROUND_UP(key_len, sizeof(u32)), hash_seed); } -struct sw_flow *flow_tbl_lookup(struct flow_table *table, +struct sw_flow *ovs_flow_tbl_lookup(struct flow_table *table, struct sw_flow_key *key, int key_len) { struct sw_flow *flow; @@ -811,7 +824,7 @@ struct sw_flow *flow_tbl_lookup(struct flow_table *table, struct hlist_head *head; u32 hash; - hash = flow_hash(key, key_len); + hash = ovs_flow_hash(key, key_len); head = find_bucket(table, hash); hlist_for_each_entry_rcu(flow, n, head, hash_node) { @@ -824,7 +837,7 @@ struct sw_flow *flow_tbl_lookup(struct flow_table *table, return NULL; } -void flow_tbl_insert(struct flow_table *table, struct sw_flow *flow) +void ovs_flow_tbl_insert(struct flow_table *table, struct sw_flow *flow) { struct hlist_head *head; @@ -833,7 +846,7 @@ void flow_tbl_insert(struct flow_table *table, struct sw_flow *flow) table->count++; } -void flow_tbl_remove(struct flow_table *table, struct sw_flow *flow) +void ovs_flow_tbl_remove(struct flow_table *table, struct sw_flow *flow) { if (!hlist_unhashed(&flow->hash_node)) { hlist_del_init_rcu(&flow->hash_node); @@ -843,11 +856,12 @@ void flow_tbl_remove(struct flow_table *table, struct sw_flow *flow) } /* The size of the argument for each %OVS_KEY_ATTR_* Netlink attribute. */ -const u32 ovs_key_lens[OVS_KEY_ATTR_MAX + 1] = { +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_ETHERNET] = sizeof(struct ovs_key_ethernet), - [OVS_KEY_ATTR_8021Q] = sizeof(struct ovs_key_8021q), + [OVS_KEY_ATTR_VLAN] = sizeof(__be16), [OVS_KEY_ATTR_ETHERTYPE] = sizeof(__be16), [OVS_KEY_ATTR_IPV4] = sizeof(struct ovs_key_ipv4), [OVS_KEY_ATTR_IPV6] = sizeof(struct ovs_key_ipv6), @@ -968,38 +982,57 @@ static int ipv6_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_len, return 0; } +static int parse_flow_nlattrs(const struct nlattr *attr, + const struct nlattr *a[], u64 *attrsp) +{ + const struct nlattr *nla; + u64 attrs; + int rem; + + attrs = 0; + nla_for_each_nested(nla, attr, rem) { + u16 type = nla_type(nla); + int expected_len; + + if (type > OVS_KEY_ATTR_MAX || attrs & (1ULL << type)) + return -EINVAL; + + expected_len = ovs_key_lens[type]; + if (nla_len(nla) != expected_len && expected_len != -1) + return -EINVAL; + + attrs |= 1ULL << type; + a[type] = nla; + } + if (rem) + return -EINVAL; + + *attrsp = attrs; + return 0; +} + /** - * flow_from_nlattrs - parses Netlink attributes into a flow key. + * 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 flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp, +int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp, const struct nlattr *attr) { const struct nlattr *a[OVS_KEY_ATTR_MAX + 1]; const struct ovs_key_ethernet *eth_key; - const struct nlattr *nla; int key_len; u64 attrs; - int rem; + int err; memset(swkey, 0, sizeof(struct sw_flow_key)); key_len = SW_FLOW_KEY_OFFSET(eth); - attrs = 0; - nla_for_each_nested(nla, attr, rem) { - u16 type = nla_type(nla); - - if (type > OVS_KEY_ATTR_MAX || attrs & (1ULL << type) || - nla_len(nla) != ovs_key_lens[type]) - return -EINVAL; - attrs |= 1ULL << type; - a[type] = nla; - } - if (rem) - return -EINVAL; + err = parse_flow_nlattrs(attr, a, &attrs); + if (err) + return err; /* Metadata attributes. */ if (attrs & (1 << OVS_KEY_ATTR_PRIORITY)) { @@ -1030,18 +1063,35 @@ int flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp, memcpy(swkey->eth.src, eth_key->eth_src, ETH_ALEN); memcpy(swkey->eth.dst, eth_key->eth_dst, ETH_ALEN); - if (attrs & (1 << OVS_KEY_ATTR_8021Q)) { - const struct ovs_key_8021q *q_key; + if (attrs & (1u << OVS_KEY_ATTR_ETHERTYPE) && + nla_get_be16(a[OVS_KEY_ATTR_ETHERTYPE]) == htons(ETH_P_8021Q)) { + const struct nlattr *encap; + __be16 tci; - q_key = nla_data(a[OVS_KEY_ATTR_8021Q]); - /* Only standard 0x8100 VLANs currently supported. */ - if (q_key->q_tpid != htons(ETH_P_8021Q)) - return -EINVAL; - if (q_key->q_tci & htons(VLAN_TAG_PRESENT)) + if (attrs != ((1 << OVS_KEY_ATTR_VLAN) | + (1 << OVS_KEY_ATTR_ETHERTYPE) | + (1 << OVS_KEY_ATTR_ENCAP))) return -EINVAL; - swkey->eth.tci = q_key->q_tci | htons(VLAN_TAG_PRESENT); - attrs &= ~(1 << OVS_KEY_ATTR_8021Q); + 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; + + 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; + + swkey->eth.type = htons(ETH_P_8021Q); + *key_lenp = key_len; + return 0; + } else { + return -EINVAL; + } } if (attrs & (1 << OVS_KEY_ATTR_ETHERTYPE)) { @@ -1072,7 +1122,7 @@ int flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp, swkey->ipv4.addr.dst = ipv4_key->ipv4_dst; if (swkey->ip.frag != OVS_FRAG_TYPE_LATER) { - int err = ipv4_flow_from_nlattrs(swkey, &key_len, a, &attrs); + err = ipv4_flow_from_nlattrs(swkey, &key_len, a, &attrs); if (err) return err; } @@ -1098,7 +1148,7 @@ int flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp, sizeof(swkey->ipv6.addr.dst)); if (swkey->ip.frag != OVS_FRAG_TYPE_LATER) { - int err = ipv6_flow_from_nlattrs(swkey, &key_len, a, &attrs); + err = ipv6_flow_from_nlattrs(swkey, &key_len, a, &attrs); if (err) return err; } @@ -1128,7 +1178,7 @@ int flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp, } /** - * flow_metadata_from_nlattrs - parses Netlink attributes into a flow key. + * 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 @@ -1139,8 +1189,8 @@ int 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 flow_metadata_from_nlattrs(u32 *priority, u16 *in_port, __be64 *tun_id, - const struct nlattr *attr) +int ovs_flow_metadata_from_nlattrs(u32 *priority, u16 *in_port, __be64 *tun_id, + const struct nlattr *attr) { const struct nlattr *nla; int rem; @@ -1152,7 +1202,7 @@ int flow_metadata_from_nlattrs(u32 *priority, u16 *in_port, __be64 *tun_id, nla_for_each_nested(nla, attr, rem) { int type = nla_type(nla); - if (type <= OVS_KEY_ATTR_MAX && ovs_key_lens[type] != 0) { + if (type <= OVS_KEY_ATTR_MAX && ovs_key_lens[type] > 0) { if (nla_len(nla) != ovs_key_lens[type]) return -EINVAL; @@ -1178,10 +1228,10 @@ int flow_metadata_from_nlattrs(u32 *priority, u16 *in_port, __be64 *tun_id, return 0; } -int flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) +int ovs_flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) { struct ovs_key_ethernet *eth_key; - struct nlattr *nla; + struct nlattr *nla, *encap; if (swkey->phy.priority) NLA_PUT_U32(skb, OVS_KEY_ATTR_PRIORITY, swkey->phy.priority); @@ -1199,16 +1249,18 @@ int flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) memcpy(eth_key->eth_src, swkey->eth.src, ETH_ALEN); memcpy(eth_key->eth_dst, swkey->eth.dst, ETH_ALEN); - if (swkey->eth.tci != htons(0)) { - struct ovs_key_8021q q_key; - - q_key.q_tpid = htons(ETH_P_8021Q); - q_key.q_tci = swkey->eth.tci & ~htons(VLAN_TAG_PRESENT); - NLA_PUT(skb, OVS_KEY_ATTR_8021Q, sizeof(q_key), &q_key); + 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); + encap = nla_nest_start(skb, OVS_KEY_ATTR_ENCAP); + if (!swkey->eth.tci) + goto unencap; + } else { + encap = NULL; } if (swkey->eth.type == htons(ETH_P_802_2)) - return 0; + goto unencap; NLA_PUT_BE16(skb, OVS_KEY_ATTR_ETHERTYPE, swkey->eth.type); @@ -1326,6 +1378,10 @@ int flow_to_nlattrs(const struct sw_flow_key *swkey, struct sk_buff *skb) } } +unencap: + if (encap) + nla_nest_end(skb, encap); + return 0; nla_put_failure: @@ -1334,7 +1390,7 @@ nla_put_failure: /* Initializes the flow module. * Returns zero if successful or a negative error code. */ -int flow_init(void) +int ovs_flow_init(void) { flow_cache = kmem_cache_create("sw_flow", sizeof(struct sw_flow), 0, 0, NULL); @@ -1347,7 +1403,7 @@ int flow_init(void) } /* Uninitializes the flow module. */ -void flow_exit(void) +void ovs_flow_exit(void) { kmem_cache_destroy(flow_cache); }