X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=datapath%2Ftunnel.c;h=19056b97b4f9c3d5a8decaff75abba7c5738128f;hb=77912ae717298cb06544f1bebd9701d0b3c2f2f0;hp=5b3d7f51d6694919d82927093fdf31d7e0c89197;hpb=681f040ea84a4283635654c4fbe6060cd6153fd8;p=openvswitch diff --git a/datapath/tunnel.c b/datapath/tunnel.c index 5b3d7f51..19056b97 100644 --- a/datapath/tunnel.c +++ b/datapath/tunnel.c @@ -10,8 +10,10 @@ #include #include #include +#include #include #include +#include #include #include #include @@ -84,8 +86,10 @@ static DECLARE_DELAYED_WORK(cache_cleaner_wq, cache_cleaner); */ static unsigned int key_local_remote_ports __read_mostly; static unsigned int key_remote_ports __read_mostly; +static unsigned int key_multicast_ports __read_mostly; static unsigned int local_remote_ports __read_mostly; static unsigned int remote_ports __read_mostly; +static unsigned int multicast_ports __read_mostly; #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,36) #define rt_dst(rt) (rt->dst) @@ -93,7 +97,20 @@ static unsigned int remote_ports __read_mostly; #define rt_dst(rt) (rt->u.dst) #endif -static inline struct vport *tnl_vport_to_vport(const struct tnl_vport *tnl_vport) +#if LINUX_VERSION_CODE >= KERNEL_VERSION(3,1,0) +static struct hh_cache *rt_hh(struct rtable *rt) +{ + struct neighbour *neigh = dst_get_neighbour(&rt->dst); + if (!neigh || !(neigh->nud_state & NUD_CONNECTED) || + !neigh->hh.hh_len) + return NULL; + return &neigh->hh; +} +#else +#define rt_hh(rt) (rt_dst(rt).hh) +#endif + +static struct vport *tnl_vport_to_vport(const struct tnl_vport *tnl_vport) { return vport_from_priv(tnl_vport); } @@ -101,13 +118,13 @@ static inline struct vport *tnl_vport_to_vport(const struct tnl_vport *tnl_vport /* This is analogous to rtnl_dereference for the tunnel cache. It checks that * cache_lock is held, so it is only for update side code. */ -static inline struct tnl_cache *cache_dereference(struct tnl_vport *tnl_vport) +static struct tnl_cache *cache_dereference(struct tnl_vport *tnl_vport) { return rcu_dereference_protected(tnl_vport->cache, - lockdep_is_held(&tnl_vport->cache_lock)); + lockdep_is_held(&tnl_vport->cache_lock)); } -static inline void schedule_cache_cleaner(void) +static void schedule_cache_cleaner(void) { schedule_delayed_work(&cache_cleaner_wq, CACHE_CLEANER_INTERVAL); } @@ -134,6 +151,21 @@ static void free_cache_rcu(struct rcu_head *rcu) free_cache(c); } +/* Frees the portion of 'mutable' that requires RTNL and thus can't happen + * within an RCU callback. Fortunately this part doesn't require waiting for + * an RCU grace period. + */ +static void free_mutable_rtnl(struct tnl_mutable_config *mutable) +{ + ASSERT_RTNL(); + if (ipv4_is_multicast(mutable->key.daddr) && mutable->mlink) { + struct in_device *in_dev; + in_dev = inetdev_by_index(&init_net, mutable->mlink); + if (in_dev) + ip_mc_dec_group(in_dev, mutable->key.daddr); + } +} + static void assign_config_rcu(struct vport *vport, struct tnl_mutable_config *new_config) { @@ -142,6 +174,8 @@ static void assign_config_rcu(struct vport *vport, old_config = rtnl_dereference(tnl_vport->mutable); rcu_assign_pointer(tnl_vport->mutable, new_config); + + free_mutable_rtnl(old_config); call_rcu(&old_config->rcu, free_config_rcu); } @@ -159,14 +193,20 @@ static void assign_cache_rcu(struct vport *vport, struct tnl_cache *new_cache) static unsigned int *find_port_pool(const struct tnl_mutable_config *mutable) { + bool is_multicast = ipv4_is_multicast(mutable->key.daddr); + if (mutable->flags & TNL_F_IN_KEY_MATCH) { if (mutable->key.saddr) return &local_remote_ports; + else if (is_multicast) + return &multicast_ports; else return &remote_ports; } else { if (mutable->key.saddr) return &key_local_remote_ports; + else if (is_multicast) + return &key_multicast_ports; else return &key_remote_ports; } @@ -174,10 +214,10 @@ static unsigned int *find_port_pool(const struct tnl_mutable_config *mutable) static u32 port_hash(const struct port_lookup_key *key) { - return jhash2((u32*)key, (PORT_KEY_LEN / sizeof(u32)), 0); + return jhash2((u32 *)key, (PORT_KEY_LEN / sizeof(u32)), 0); } -static inline struct hlist_head *find_bucket(u32 hash) +static struct hlist_head *find_bucket(u32 hash) { return &port_table[(hash & (PORT_TABLE_SIZE - 1))]; } @@ -233,7 +273,7 @@ static struct vport *port_table_lookup(struct port_lookup_key *key, struct hlist_node *n; struct hlist_head *bucket; u32 hash = port_hash(key); - struct tnl_vport * tnl_vport; + struct tnl_vport *tnl_vport; bucket = find_bucket(hash); @@ -256,42 +296,56 @@ struct vport *tnl_find_port(__be32 saddr, __be32 daddr, __be64 key, { struct port_lookup_key lookup; struct vport *vport; + bool is_multicast = ipv4_is_multicast(saddr); lookup.saddr = saddr; lookup.daddr = daddr; - if (tunnel_type & TNL_T_KEY_EXACT) { - lookup.in_key = key; - lookup.tunnel_type = tunnel_type & ~TNL_T_KEY_MATCH; - - if (key_local_remote_ports) { - vport = port_table_lookup(&lookup, mutable); - if (vport) - return vport; - } - - if (key_remote_ports) { - lookup.saddr = 0; - vport = port_table_lookup(&lookup, mutable); - if (vport) - return vport; + /* First try for exact match on in_key. */ + lookup.in_key = key; + lookup.tunnel_type = tunnel_type | TNL_T_KEY_EXACT; + if (!is_multicast && key_local_remote_ports) { + vport = port_table_lookup(&lookup, mutable); + if (vport) + return vport; + } + if (key_remote_ports) { + lookup.saddr = 0; + vport = port_table_lookup(&lookup, mutable); + if (vport) + return vport; - lookup.saddr = saddr; - } + lookup.saddr = saddr; } - if (tunnel_type & TNL_T_KEY_MATCH) { - lookup.in_key = 0; - lookup.tunnel_type = tunnel_type & ~TNL_T_KEY_EXACT; + /* Then try matches that wildcard in_key. */ + lookup.in_key = 0; + lookup.tunnel_type = tunnel_type | TNL_T_KEY_MATCH; + if (!is_multicast && local_remote_ports) { + vport = port_table_lookup(&lookup, mutable); + if (vport) + return vport; + } + if (remote_ports) { + lookup.saddr = 0; + vport = port_table_lookup(&lookup, mutable); + if (vport) + return vport; + } - if (local_remote_ports) { + if (is_multicast) { + lookup.saddr = 0; + lookup.daddr = saddr; + if (key_multicast_ports) { + lookup.tunnel_type = tunnel_type | TNL_T_KEY_EXACT; + lookup.in_key = key; vport = port_table_lookup(&lookup, mutable); if (vport) return vport; } - - if (remote_ports) { - lookup.saddr = 0; + if (multicast_ports) { + lookup.tunnel_type = tunnel_type | TNL_T_KEY_MATCH; + lookup.in_key = 0; vport = port_table_lookup(&lookup, mutable); if (vport) return vport; @@ -544,7 +598,8 @@ static void ipv6_build_icmp(struct sk_buff *skb, struct sk_buff *nskb, } #endif /* IPv6 */ -bool tnl_frag_needed(struct vport *vport, const struct tnl_mutable_config *mutable, +bool tnl_frag_needed(struct vport *vport, + const struct tnl_mutable_config *mutable, struct sk_buff *skb, unsigned int mtu, __be64 flow_key) { unsigned int eth_hdr_len = ETH_HLEN; @@ -741,15 +796,21 @@ static void create_tunnel_header(const struct vport *vport, tnl_vport->tnl_ops->build_header(vport, mutable, iph + 1); } -static inline void *get_cached_header(const struct tnl_cache *cache) +static void *get_cached_header(const struct tnl_cache *cache) { return (void *)cache + ALIGN(sizeof(struct tnl_cache), CACHE_DATA_ALIGN); } -static inline bool check_cache_valid(const struct tnl_cache *cache, - const struct tnl_mutable_config *mutable) +static bool check_cache_valid(const struct tnl_cache *cache, + const struct tnl_mutable_config *mutable) { - return cache && + struct hh_cache *hh; + + if (!cache) + return false; + + hh = rt_hh(cache->rt); + return hh && #ifdef NEED_CACHE_TIMEOUT time_before(jiffies, cache->expiration) && #endif @@ -757,7 +818,7 @@ static inline bool check_cache_valid(const struct tnl_cache *cache, atomic_read(&init_net.ipv4.rt_genid) == cache->rt->rt_genid && #endif #ifdef HAVE_HH_SEQ - rt_dst(cache->rt).hh->hh_lock.sequence == cache->hh_seq && + hh->hh_lock.sequence == cache->hh_seq && #endif mutable->seq == cache->mutable_seq && (!is_internal_dev(rt_dst(cache->rt).dev) || @@ -796,26 +857,28 @@ static void cache_cleaner(struct work_struct *work) rcu_read_unlock(); } -static inline void create_eth_hdr(struct tnl_cache *cache, - const struct rtable *rt) +static void create_eth_hdr(struct tnl_cache *cache, struct hh_cache *hh) { void *cache_data = get_cached_header(cache); - int hh_len = rt_dst(rt).hh->hh_len; - int hh_off = HH_DATA_ALIGN(rt_dst(rt).hh->hh_len) - hh_len; + int hh_off; #ifdef HAVE_HH_SEQ unsigned hh_seq; do { - hh_seq = read_seqbegin(&rt_dst(rt).hh->hh_lock); - memcpy(cache_data, (void *)rt_dst(rt).hh->hh_data + hh_off, hh_len); - } while (read_seqretry(&rt_dst(rt).hh->hh_lock, hh_seq)); + hh_seq = read_seqbegin(&hh->hh_lock); + hh_off = HH_DATA_ALIGN(hh->hh_len) - hh->hh_len; + memcpy(cache_data, (void *)hh->hh_data + hh_off, hh->hh_len); + cache->hh_len = hh->hh_len; + } while (read_seqretry(&hh->hh_lock, hh_seq)); cache->hh_seq = hh_seq; #else - read_lock(&rt_dst(rt).hh->hh_lock); - memcpy(cache_data, (void *)rt_dst(rt).hh->hh_data + hh_off, hh_len); - read_unlock(&rt_dst(rt).hh->hh_lock); + read_lock(&hh->hh_lock); + hh_off = HH_DATA_ALIGN(hh->hh_len) - hh->hh_len; + memcpy(cache_data, (void *)hh->hh_data + hh_off, hh->hh_len); + cache->hh_len = hh->hh_len; + read_unlock(&hh->hh_lock); #endif } @@ -827,6 +890,7 @@ static struct tnl_cache *build_cache(struct vport *vport, struct tnl_cache *cache; void *cache_data; int cache_len; + struct hh_cache *hh; if (!(mutable->flags & TNL_F_HDR_CACHE)) return NULL; @@ -835,7 +899,9 @@ static struct tnl_cache *build_cache(struct vport *vport, * If there is no entry in the ARP cache or if this device does not * support hard header caching just fall back to the IP stack. */ - if (!rt_dst(rt).hh) + + hh = rt_hh(rt); + if (!hh) return NULL; /* @@ -851,17 +917,16 @@ static struct tnl_cache *build_cache(struct vport *vport, else cache = NULL; - cache_len = rt_dst(rt).hh->hh_len + mutable->tunnel_hlen; + cache_len = LL_RESERVED_SPACE(rt_dst(rt).dev) + mutable->tunnel_hlen; cache = kzalloc(ALIGN(sizeof(struct tnl_cache), CACHE_DATA_ALIGN) + cache_len, GFP_ATOMIC); if (!cache) goto unlock; - cache->len = cache_len; - - create_eth_hdr(cache, rt); - cache_data = get_cached_header(cache) + rt_dst(rt).hh->hh_len; + create_eth_hdr(cache, hh); + cache_data = get_cached_header(cache) + cache->hh_len; + cache->len = cache->hh_len + mutable->tunnel_hlen; create_tunnel_header(vport, mutable, rt, cache_data); @@ -875,7 +940,6 @@ static struct tnl_cache *build_cache(struct vport *vport, struct sw_flow_key flow_key; struct vport *dst_vport; struct sk_buff *skb; - bool is_frag; int err; int flow_key_len; struct sw_flow *flow; @@ -892,10 +956,10 @@ static struct tnl_cache *build_cache(struct vport *vport, memcpy(skb->data, get_cached_header(cache), cache->len); err = flow_extract(skb, dst_vport->port_no, &flow_key, - &flow_key_len, &is_frag); + &flow_key_len); consume_skb(skb); - if (err || is_frag) + if (err) goto done; flow = flow_tbl_lookup(rcu_dereference(dst_vport->dp->table), @@ -915,6 +979,31 @@ unlock: return cache; } +static struct rtable *__find_route(const struct tnl_mutable_config *mutable, + u8 ipproto, u8 tos) +{ +#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,39) + struct flowi fl = { .nl_u = { .ip4_u = { + .daddr = mutable->key.daddr, + .saddr = mutable->key.saddr, + .tos = tos } }, + .proto = ipproto }; + struct rtable *rt; + + if (unlikely(ip_route_output_key(&init_net, &rt, &fl))) + return ERR_PTR(-EADDRNOTAVAIL); + + return rt; +#else + struct flowi4 fl = { .daddr = mutable->key.daddr, + .saddr = mutable->key.saddr, + .flowi4_tos = tos, + .flowi4_proto = ipproto }; + + return ip_route_output_key(&init_net, &fl); +#endif +} + static struct rtable *find_route(struct vport *vport, const struct tnl_mutable_config *mutable, u8 tos, struct tnl_cache **cache) @@ -925,30 +1014,16 @@ static struct rtable *find_route(struct vport *vport, *cache = NULL; tos = RT_TOS(tos); - if (likely(tos == mutable->tos && check_cache_valid(cur_cache, mutable))) { + if (likely(tos == mutable->tos && + check_cache_valid(cur_cache, mutable))) { *cache = cur_cache; return cur_cache->rt; } else { struct rtable *rt; -#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,39) - struct flowi fl = { .nl_u = { .ip4_u = - { .daddr = mutable->key.daddr, - .saddr = mutable->key.saddr, - .tos = tos } }, - .proto = tnl_vport->tnl_ops->ipproto }; - if (unlikely(ip_route_output_key(&init_net, &rt, &fl))) - return NULL; -#else - struct flowi4 fl = { .daddr = mutable->key.daddr, - .saddr = mutable->key.saddr, - .flowi4_tos = tos, - .flowi4_proto = tnl_vport->tnl_ops->ipproto }; - - rt = ip_route_output_key(&init_net, &fl); + rt = __find_route(mutable, tnl_vport->tnl_ops->ipproto, tos); if (IS_ERR(rt)) return NULL; -#endif if (likely(tos == mutable->tos)) *cache = build_cache(vport, mutable, rt); @@ -957,7 +1032,7 @@ static struct rtable *find_route(struct vport *vport, } } -static inline bool need_linearize(const struct sk_buff *skb) +static bool need_linearize(const struct sk_buff *skb) { int i; @@ -970,7 +1045,7 @@ static inline bool need_linearize(const struct sk_buff *skb) * change them from underneath us and we can skip the linearization. */ for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) - if (unlikely(page_count(skb_shinfo(skb)->frags[i].page) > 1)) + if (unlikely(page_count(skb_frag_page(&skb_shinfo(skb)->frags[i])) > 1)) return true; return false; @@ -1187,7 +1262,7 @@ int tnl_send(struct vport *vport, struct sk_buff *skb) skb_push(skb, cache->len); memcpy(skb->data, get_cached_header(cache), cache->len); skb_reset_mac_header(skb); - skb_set_network_header(skb, rt_dst(rt).hh->hh_len); + skb_set_network_header(skb, cache->hh_len); } else { skb_push(skb, mutable->tunnel_hlen); @@ -1209,14 +1284,16 @@ int tnl_send(struct vport *vport, struct sk_buff *skb) iph->frag_off = frag_off; ip_select_ident(iph, &rt_dst(rt), NULL); - skb = tnl_vport->tnl_ops->update_header(vport, mutable, &rt_dst(rt), skb); + skb = tnl_vport->tnl_ops->update_header(vport, mutable, + &rt_dst(rt), skb); if (unlikely(!skb)) goto next; if (likely(cache)) { int orig_len = skb->len - cache->len; - struct vport *cache_vport = internal_dev_get_vport(rt_dst(rt).dev); + struct vport *cache_vport; + cache_vport = internal_dev_get_vport(rt_dst(rt).dev); skb->protocol = htons(ETH_P_IP); iph = ip_hdr(skb); iph->tot_len = htons(skb->len - skb_network_offset(skb)); @@ -1271,7 +1348,8 @@ static const struct nla_policy tnl_policy[OVS_TUNNEL_ATTR_MAX + 1] = { [OVS_TUNNEL_ATTR_TTL] = { .type = NLA_U8 }, }; -/* Sets OVS_TUNNEL_ATTR_* fields in 'mutable', which must initially be zeroed. */ +/* Sets OVS_TUNNEL_ATTR_* fields in 'mutable', which must initially be + * zeroed. */ static int tnl_set_config(struct nlattr *options, const struct tnl_ops *tnl_ops, const struct vport *cur_vport, struct tnl_mutable_config *mutable) @@ -1293,9 +1371,12 @@ static int tnl_set_config(struct nlattr *options, const struct tnl_ops *tnl_ops, mutable->flags = nla_get_u32(a[OVS_TUNNEL_ATTR_FLAGS]) & TNL_F_PUBLIC; - if (a[OVS_TUNNEL_ATTR_SRC_IPV4]) - mutable->key.saddr = nla_get_be32(a[OVS_TUNNEL_ATTR_SRC_IPV4]); mutable->key.daddr = nla_get_be32(a[OVS_TUNNEL_ATTR_DST_IPV4]); + if (a[OVS_TUNNEL_ATTR_SRC_IPV4]) { + if (ipv4_is_multicast(mutable->key.daddr)) + return -EINVAL; + mutable->key.saddr = nla_get_be32(a[OVS_TUNNEL_ATTR_SRC_IPV4]); + } if (a[OVS_TUNNEL_ATTR_TOS]) { mutable->tos = nla_get_u8(a[OVS_TUNNEL_ATTR_TOS]); @@ -1326,13 +1407,26 @@ static int tnl_set_config(struct nlattr *options, const struct tnl_ops *tnl_ops, mutable->tunnel_hlen += sizeof(struct iphdr); - old_vport = tnl_find_port(mutable->key.saddr, mutable->key.daddr, - mutable->key.in_key, mutable->key.tunnel_type, - &old_mutable); - + old_vport = port_table_lookup(&mutable->key, &old_mutable); if (old_vport && old_vport != cur_vport) return -EEXIST; + mutable->mlink = 0; + if (ipv4_is_multicast(mutable->key.daddr)) { + struct net_device *dev; + struct rtable *rt; + + rt = __find_route(mutable, tnl_ops->ipproto, mutable->tos); + if (IS_ERR(rt)) + return -EADDRNOTAVAIL; + dev = rt_dst(rt).dev; + ip_rt_put(rt); + if (__in_dev_get_rtnl(dev) == NULL) + return -EADDRNOTAVAIL; + mutable->mlink = dev->ifindex; + ip_mc_inc_group(__in_dev_get_rtnl(dev), mutable->key.daddr); + } + return 0; } @@ -1385,6 +1479,7 @@ struct vport *tnl_create(const struct vport_parms *parms, return vport; error_free_mutable: + free_mutable_rtnl(mutable); kfree(mutable); error_free_vport: vport_free(vport); @@ -1417,10 +1512,13 @@ int tnl_set_options(struct vport *vport, struct nlattr *options) if (port_hash(&mutable->key) != port_hash(&old_mutable->key)) port_table_move_port(vport, mutable); + else + assign_config_rcu(vport, mutable); return 0; error_free: + free_mutable_rtnl(mutable); kfree(mutable); error: return err; @@ -1464,23 +1562,26 @@ static void free_port_rcu(struct rcu_head *rcu) void tnl_destroy(struct vport *vport) { struct tnl_vport *tnl_vport = tnl_vport_priv(vport); - const struct tnl_mutable_config *mutable; + struct tnl_mutable_config *mutable; mutable = rtnl_dereference(tnl_vport->mutable); port_table_remove_port(vport); + free_mutable_rtnl(mutable); call_rcu(&tnl_vport->rcu, free_port_rcu); } int tnl_set_addr(struct vport *vport, const unsigned char *addr) { struct tnl_vport *tnl_vport = tnl_vport_priv(vport); - struct tnl_mutable_config *mutable; + struct tnl_mutable_config *old_mutable, *mutable; - mutable = kmemdup(rtnl_dereference(tnl_vport->mutable), - sizeof(struct tnl_mutable_config), GFP_KERNEL); + old_mutable = rtnl_dereference(tnl_vport->mutable); + mutable = kmemdup(old_mutable, sizeof(struct tnl_mutable_config), GFP_KERNEL); if (!mutable) return -ENOMEM; + old_mutable->mlink = 0; + memcpy(mutable->eth_addr, addr, ETH_ALEN); assign_config_rcu(vport, mutable); @@ -1528,7 +1629,7 @@ void tnl_exit(void) int i; for (i = 0; i < PORT_TABLE_SIZE; i++) { - struct tnl_vport * tnl_vport; + struct tnl_vport *tnl_vport; struct hlist_head *hash_head; struct hlist_node *n;