X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=datapath%2Ftunnel.c;h=d651c11833885a573cc743d1a08bc6d2a40e001e;hb=5ca92d1d5dc0d8dba9ed554444cd0ae00a43209f;hp=a6fdfa94e65b1255f89b72e706657958654163d4;hpb=ae05bcfd103fcd67953ad86cbd541288014c3443;p=openvswitch diff --git a/datapath/tunnel.c b/datapath/tunnel.c index a6fdfa94..d651c118 100644 --- a/datapath/tunnel.c +++ b/datapath/tunnel.c @@ -1,11 +1,23 @@ /* - * Copyright (c) 2010, 2011 Nicira Networks. - * Distributed under the terms of the GNU GPL version 2. + * Copyright (c) 2007-2012 Nicira, Inc. * - * 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 */ +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + #include #include #include @@ -32,7 +44,6 @@ #include #include -#include "actions.h" #include "checksum.h" #include "datapath.h" #include "tunnel.h" @@ -100,7 +111,7 @@ static unsigned int multicast_ports __read_mostly; #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); + struct neighbour *neigh = dst_get_neighbour_noref(&rt->dst); if (!neigh || !(neigh->nud_state & NUD_CONNECTED) || !neigh->hh.hh_len) return NULL; @@ -134,7 +145,7 @@ static void free_cache(struct tnl_cache *cache) if (!cache) return; - flow_put(cache->flow); + ovs_flow_put(cache->flow); ip_rt_put(cache->rt); kfree(cache); } @@ -160,7 +171,7 @@ 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); + in_dev = inetdev_by_index(port_key_get_net(&mutable->key), mutable->mlink); if (in_dev) ip_mc_dec_group(in_dev, mutable->key.daddr); } @@ -290,14 +301,15 @@ static struct vport *port_table_lookup(struct port_lookup_key *key, return NULL; } -struct vport *tnl_find_port(__be32 saddr, __be32 daddr, __be64 key, - int tunnel_type, - const struct tnl_mutable_config **mutable) +struct vport *ovs_tnl_find_port(struct net *net, __be32 saddr, __be32 daddr, + __be64 key, int tunnel_type, + const struct tnl_mutable_config **mutable) { struct port_lookup_key lookup; struct vport *vport; bool is_multicast = ipv4_is_multicast(saddr); + port_key_set_net(&lookup, net); lookup.saddr = saddr; lookup.daddr = daddr; @@ -390,7 +402,7 @@ static void ecn_decapsulate(struct sk_buff *skb, u8 tos) } /** - * tnl_rcv - ingress point for generic tunnel code + * ovs_tnl_rcv - ingress point for generic tunnel code * * @vport: port this packet was received on * @skb: received packet @@ -404,7 +416,7 @@ static void ecn_decapsulate(struct sk_buff *skb, u8 tos) * - skb->csum does not include the inner Ethernet header. * - The layer pointers are undefined. */ -void tnl_rcv(struct vport *vport, struct sk_buff *skb, u8 tos) +void ovs_tnl_rcv(struct vport *vport, struct sk_buff *skb, u8 tos) { struct ethhdr *eh; @@ -429,7 +441,7 @@ void tnl_rcv(struct vport *vport, struct sk_buff *skb, u8 tos) return; } - vport_receive(vport, skb); + ovs_vport_receive(vport, skb); } static bool check_ipv4_address(__be32 addr) @@ -529,6 +541,7 @@ static bool ipv6_should_icmp(struct sk_buff *skb) int addr_type; int payload_off = (u8 *)(old_ipv6h + 1) - skb->data; u8 nexthdr = ipv6_hdr(skb)->nexthdr; + __be16 frag_off; /* Check source address is valid. */ addr_type = ipv6_addr_type(&old_ipv6h->saddr); @@ -540,7 +553,7 @@ static bool ipv6_should_icmp(struct sk_buff *skb) return false; /* Don't respond to ICMP error messages. */ - payload_off = ipv6_skip_exthdr(skb, payload_off, &nexthdr); + payload_off = ipv6_skip_exthdr(skb, payload_off, &nexthdr, &frag_off); if (payload_off < 0) return false; @@ -578,8 +591,8 @@ static void ipv6_build_icmp(struct sk_buff *skb, struct sk_buff *nskb, + payload_length); ipv6h->nexthdr = NEXTHDR_ICMP; ipv6h->hop_limit = IPV6_DEFAULT_HOPLIMIT; - ipv6_addr_copy(&ipv6h->daddr, &old_ipv6h->saddr); - ipv6_addr_copy(&ipv6h->saddr, &old_ipv6h->daddr); + ipv6h->daddr = old_ipv6h->saddr; + ipv6h->saddr = old_ipv6h->daddr; /* ICMPv6 */ icmp6h->icmp6_type = ICMPV6_PKT_TOOBIG; @@ -598,9 +611,9 @@ 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, - struct sk_buff *skb, unsigned int mtu, __be64 flow_key) +bool ovs_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; unsigned int total_length = 0, header_length = 0, payload_length; @@ -700,7 +713,7 @@ bool tnl_frag_needed(struct vport *vport, return false; } - vport_receive(vport, nskb); + ovs_vport_receive(vport, nskb); return true; } @@ -747,8 +760,8 @@ static bool check_mtu(struct sk_buff *skb, mtu = max(mtu, IP_MIN_MTU); if (packet_length > mtu && - tnl_frag_needed(vport, mutable, skb, mtu, - OVS_CB(skb)->tun_id)) + ovs_tnl_frag_needed(vport, mutable, skb, mtu, + OVS_CB(skb)->tun_id)) return false; } } @@ -764,8 +777,8 @@ static bool check_mtu(struct sk_buff *skb, mtu = max(mtu, IPV6_MIN_MTU); if (packet_length > mtu && - tnl_frag_needed(vport, mutable, skb, mtu, - OVS_CB(skb)->tun_id)) + ovs_tnl_frag_needed(vport, mutable, skb, mtu, + OVS_CB(skb)->tun_id)) return false; } } @@ -801,6 +814,13 @@ static void *get_cached_header(const struct tnl_cache *cache) return (void *)cache + ALIGN(sizeof(struct tnl_cache), CACHE_DATA_ALIGN); } +#ifdef HAVE_RT_GENID +static inline int rt_genid(struct net *net) +{ + return atomic_read(&net->ipv4.rt_genid); +} +#endif + static bool check_cache_valid(const struct tnl_cache *cache, const struct tnl_mutable_config *mutable) { @@ -815,13 +835,13 @@ static bool check_cache_valid(const struct tnl_cache *cache, time_before(jiffies, cache->expiration) && #endif #ifdef HAVE_RT_GENID - atomic_read(&init_net.ipv4.rt_genid) == cache->rt->rt_genid && + rt_genid(dev_net(rt_dst(cache->rt).dev)) == cache->rt->rt_genid && #endif #ifdef HAVE_HH_SEQ hh->hh_lock.sequence == cache->hh_seq && #endif mutable->seq == cache->mutable_seq && - (!is_internal_dev(rt_dst(cache->rt).dev) || + (!ovs_is_internal_dev(rt_dst(cache->rt).dev) || (cache->flow && !cache->flow->dead)); } @@ -848,7 +868,7 @@ static void cache_cleaner(struct work_struct *work) for (i = 0; i < PORT_TABLE_SIZE; i++) { struct hlist_node *n; struct hlist_head *bucket; - struct tnl_vport *tnl_vport; + struct tnl_vport *tnl_vport; bucket = &port_table[i]; hlist_for_each_entry_rcu(tnl_vport, n, bucket, hash_node) @@ -936,7 +956,7 @@ static struct tnl_cache *build_cache(struct vport *vport, cache->expiration = jiffies + tnl_vport->cache_exp_interval; #endif - if (is_internal_dev(rt_dst(rt).dev)) { + if (ovs_is_internal_dev(rt_dst(rt).dev)) { struct sw_flow_key flow_key; struct vport *dst_vport; struct sk_buff *skb; @@ -944,7 +964,7 @@ static struct tnl_cache *build_cache(struct vport *vport, int flow_key_len; struct sw_flow *flow; - dst_vport = internal_dev_get_vport(rt_dst(rt).dev); + dst_vport = ovs_internal_dev_get_vport(rt_dst(rt).dev); if (!dst_vport) goto done; @@ -955,18 +975,18 @@ static struct tnl_cache *build_cache(struct vport *vport, __skb_put(skb, cache->len); memcpy(skb->data, get_cached_header(cache), cache->len); - err = flow_extract(skb, dst_vport->port_no, &flow_key, - &flow_key_len); + err = ovs_flow_extract(skb, dst_vport->port_no, &flow_key, + &flow_key_len); consume_skb(skb); if (err) goto done; - flow = flow_tbl_lookup(rcu_dereference(dst_vport->dp->table), - &flow_key, flow_key_len); + flow = ovs_flow_tbl_lookup(rcu_dereference(dst_vport->dp->table), + &flow_key, flow_key_len); if (flow) { cache->flow = flow; - flow_hold(flow); + ovs_flow_hold(flow); } } @@ -982,25 +1002,28 @@ unlock: static struct rtable *__find_route(const struct tnl_mutable_config *mutable, u8 ipproto, u8 tos) { + /* Tunnel configuration keeps DSCP part of TOS bits, But Linux + * router expect RT_TOS bits only. */ + #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 }; + .tos = RT_TOS(tos) } }, + .proto = ipproto }; struct rtable *rt; - if (unlikely(ip_route_output_key(&init_net, &rt, &fl))) + if (unlikely(ip_route_output_key(port_key_get_net(&mutable->key), &rt, &fl))) return ERR_PTR(-EADDRNOTAVAIL); return rt; #else struct flowi4 fl = { .daddr = mutable->key.daddr, .saddr = mutable->key.saddr, - .flowi4_tos = tos, + .flowi4_tos = RT_TOS(tos), .flowi4_proto = ipproto }; - return ip_route_output_key(&init_net, &fl); + return ip_route_output_key(port_key_get_net(&mutable->key), &fl); #endif } @@ -1014,7 +1037,7 @@ static struct rtable *find_route(struct vport *vport, *cache = NULL; tos = RT_TOS(tos); - if (likely(tos == mutable->tos && + if (likely(tos == RT_TOS(mutable->tos) && check_cache_valid(cur_cache, mutable))) { *cache = cur_cache; return cur_cache->rt; @@ -1025,7 +1048,7 @@ static struct rtable *find_route(struct vport *vport, if (IS_ERR(rt)) return NULL; - if (likely(tos == mutable->tos)) + if (likely(tos == RT_TOS(mutable->tos))) *cache = build_cache(vport, mutable, rt); return rt; @@ -1142,11 +1165,11 @@ free_frags: * dropped so just free the rest. This may help improve the congestion * that caused the first packet to be dropped. */ - tnl_free_linked_skbs(skb); + ovs_tnl_free_linked_skbs(skb); return sent_len; } -int tnl_send(struct vport *vport, struct sk_buff *skb) +int ovs_tnl_send(struct vport *vport, struct sk_buff *skb) { struct tnl_vport *tnl_vport = tnl_vport_priv(vport); const struct tnl_mutable_config *mutable = rcu_dereference(tnl_vport->mutable); @@ -1199,8 +1222,6 @@ int tnl_send(struct vport *vport, struct sk_buff *skb) else tos = mutable->tos; - tos = INET_ECN_encapsulate(tos, inner_tos); - /* Route lookup */ rt = find_route(vport, mutable, tos, &cache); if (unlikely(!rt)) @@ -1208,6 +1229,8 @@ int tnl_send(struct vport *vport, struct sk_buff *skb) if (unlikely(!cache)) unattached_dst = &rt_dst(rt); + tos = INET_ECN_encapsulate(tos, inner_tos); + /* Reset SKB */ nf_reset(skb); secpath_reset(skb); @@ -1293,7 +1316,7 @@ int tnl_send(struct vport *vport, struct sk_buff *skb) int orig_len = skb->len - cache->len; struct vport *cache_vport; - cache_vport = internal_dev_get_vport(rt_dst(rt).dev); + cache_vport = ovs_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)); @@ -1306,7 +1329,7 @@ int tnl_send(struct vport *vport, struct sk_buff *skb) } OVS_CB(skb)->flow = cache->flow; - vport_receive(cache_vport, skb); + ovs_vport_receive(cache_vport, skb); sent_len += orig_len; } else { int xmit_err; @@ -1325,14 +1348,14 @@ next: } if (unlikely(sent_len == 0)) - vport_record_error(vport, VPORT_E_TX_DROPPED); + ovs_vport_record_error(vport, VPORT_E_TX_DROPPED); goto out; error_free: - tnl_free_linked_skbs(skb); + ovs_tnl_free_linked_skbs(skb); error: - vport_record_error(vport, err); + ovs_vport_record_error(vport, err); out: dst_release(unattached_dst); return sent_len; @@ -1350,7 +1373,8 @@ static const struct nla_policy tnl_policy[OVS_TUNNEL_ATTR_MAX + 1] = { /* 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, +static int tnl_set_config(struct net *net, struct nlattr *options, + const struct tnl_ops *tnl_ops, const struct vport *cur_vport, struct tnl_mutable_config *mutable) { @@ -1371,6 +1395,7 @@ 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; + port_key_set_net(&mutable->key, net); 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)) @@ -1380,7 +1405,8 @@ static int tnl_set_config(struct nlattr *options, const struct tnl_ops *tnl_ops, if (a[OVS_TUNNEL_ATTR_TOS]) { mutable->tos = nla_get_u8(a[OVS_TUNNEL_ATTR_TOS]); - if (mutable->tos != RT_TOS(mutable->tos)) + /* Reject ToS config with ECN bits set. */ + if (mutable->tos & INET_ECN_MASK) return -EINVAL; } @@ -1430,9 +1456,9 @@ static int tnl_set_config(struct nlattr *options, const struct tnl_ops *tnl_ops, return 0; } -struct vport *tnl_create(const struct vport_parms *parms, - const struct vport_ops *vport_ops, - const struct tnl_ops *tnl_ops) +struct vport *ovs_tnl_create(const struct vport_parms *parms, + const struct vport_ops *vport_ops, + const struct tnl_ops *tnl_ops) { struct vport *vport; struct tnl_vport *tnl_vport; @@ -1440,7 +1466,7 @@ struct vport *tnl_create(const struct vport_parms *parms, int initial_frag_id; int err; - vport = vport_alloc(sizeof(struct tnl_vport), vport_ops, parms); + vport = ovs_vport_alloc(sizeof(struct tnl_vport), vport_ops, parms); if (IS_ERR(vport)) { err = PTR_ERR(vport); goto error; @@ -1462,7 +1488,8 @@ struct vport *tnl_create(const struct vport_parms *parms, get_random_bytes(&initial_frag_id, sizeof(int)); atomic_set(&tnl_vport->frag_id, initial_frag_id); - err = tnl_set_config(parms->options, tnl_ops, NULL, mutable); + err = tnl_set_config(ovs_dp_get_net(parms->dp), parms->options, tnl_ops, + NULL, mutable); if (err) goto error_free_mutable; @@ -1482,12 +1509,12 @@ error_free_mutable: free_mutable_rtnl(mutable); kfree(mutable); error_free_vport: - vport_free(vport); + ovs_vport_free(vport); error: return ERR_PTR(err); } -int tnl_set_options(struct vport *vport, struct nlattr *options) +int ovs_tnl_set_options(struct vport *vport, struct nlattr *options) { struct tnl_vport *tnl_vport = tnl_vport_priv(vport); const struct tnl_mutable_config *old_mutable; @@ -1506,7 +1533,8 @@ int tnl_set_options(struct vport *vport, struct nlattr *options) memcpy(mutable->eth_addr, old_mutable->eth_addr, ETH_ALEN); /* Parse the others configured by userspace. */ - err = tnl_set_config(options, tnl_vport->tnl_ops, vport, mutable); + err = tnl_set_config(ovs_dp_get_net(vport->dp), options, tnl_vport->tnl_ops, + vport, mutable); if (err) goto error_free; @@ -1524,24 +1552,29 @@ error: return err; } -int tnl_get_options(const struct vport *vport, struct sk_buff *skb) +int ovs_tnl_get_options(const struct vport *vport, struct sk_buff *skb) { const struct tnl_vport *tnl_vport = tnl_vport_priv(vport); const struct tnl_mutable_config *mutable = rcu_dereference_rtnl(tnl_vport->mutable); - NLA_PUT_U32(skb, OVS_TUNNEL_ATTR_FLAGS, mutable->flags & TNL_F_PUBLIC); - NLA_PUT_BE32(skb, OVS_TUNNEL_ATTR_DST_IPV4, mutable->key.daddr); - - if (!(mutable->flags & TNL_F_IN_KEY_MATCH)) - NLA_PUT_BE64(skb, OVS_TUNNEL_ATTR_IN_KEY, mutable->key.in_key); - if (!(mutable->flags & TNL_F_OUT_KEY_ACTION)) - NLA_PUT_BE64(skb, OVS_TUNNEL_ATTR_OUT_KEY, mutable->out_key); - if (mutable->key.saddr) - NLA_PUT_BE32(skb, OVS_TUNNEL_ATTR_SRC_IPV4, mutable->key.saddr); - if (mutable->tos) - NLA_PUT_U8(skb, OVS_TUNNEL_ATTR_TOS, mutable->tos); - if (mutable->ttl) - NLA_PUT_U8(skb, OVS_TUNNEL_ATTR_TTL, mutable->ttl); + if (nla_put_u32(skb, OVS_TUNNEL_ATTR_FLAGS, + mutable->flags & TNL_F_PUBLIC) || + nla_put_be32(skb, OVS_TUNNEL_ATTR_DST_IPV4, mutable->key.daddr)) + goto nla_put_failure; + + if (!(mutable->flags & TNL_F_IN_KEY_MATCH) && + nla_put_be64(skb, OVS_TUNNEL_ATTR_IN_KEY, mutable->key.in_key)) + goto nla_put_failure; + if (!(mutable->flags & TNL_F_OUT_KEY_ACTION) && + nla_put_be64(skb, OVS_TUNNEL_ATTR_OUT_KEY, mutable->out_key)) + goto nla_put_failure; + if (mutable->key.saddr && + nla_put_be32(skb, OVS_TUNNEL_ATTR_SRC_IPV4, mutable->key.saddr)) + goto nla_put_failure; + if (mutable->tos && nla_put_u8(skb, OVS_TUNNEL_ATTR_TOS, mutable->tos)) + goto nla_put_failure; + if (mutable->ttl && nla_put_u8(skb, OVS_TUNNEL_ATTR_TTL, mutable->ttl)) + goto nla_put_failure; return 0; @@ -1556,10 +1589,10 @@ static void free_port_rcu(struct rcu_head *rcu) free_cache((struct tnl_cache __force *)tnl_vport->cache); kfree((struct tnl_mutable __force *)tnl_vport->mutable); - vport_free(tnl_vport_to_vport(tnl_vport)); + ovs_vport_free(tnl_vport_to_vport(tnl_vport)); } -void tnl_destroy(struct vport *vport) +void ovs_tnl_destroy(struct vport *vport) { struct tnl_vport *tnl_vport = tnl_vport_priv(vport); struct tnl_mutable_config *mutable; @@ -1570,7 +1603,7 @@ void tnl_destroy(struct vport *vport) call_rcu(&tnl_vport->rcu, free_port_rcu); } -int tnl_set_addr(struct vport *vport, const unsigned char *addr) +int ovs_tnl_set_addr(struct vport *vport, const unsigned char *addr) { struct tnl_vport *tnl_vport = tnl_vport_priv(vport); struct tnl_mutable_config *old_mutable, *mutable; @@ -1588,19 +1621,19 @@ int tnl_set_addr(struct vport *vport, const unsigned char *addr) return 0; } -const char *tnl_get_name(const struct vport *vport) +const char *ovs_tnl_get_name(const struct vport *vport) { const struct tnl_vport *tnl_vport = tnl_vport_priv(vport); return tnl_vport->name; } -const unsigned char *tnl_get_addr(const struct vport *vport) +const unsigned char *ovs_tnl_get_addr(const struct vport *vport) { const struct tnl_vport *tnl_vport = tnl_vport_priv(vport); return rcu_dereference_rtnl(tnl_vport->mutable)->eth_addr; } -void tnl_free_linked_skbs(struct sk_buff *skb) +void ovs_tnl_free_linked_skbs(struct sk_buff *skb) { while (skb) { struct sk_buff *next = skb->next; @@ -1609,12 +1642,12 @@ void tnl_free_linked_skbs(struct sk_buff *skb) } } -int tnl_init(void) +int ovs_tnl_init(void) { int i; port_table = kmalloc(PORT_TABLE_SIZE * sizeof(struct hlist_head *), - GFP_KERNEL); + GFP_KERNEL); if (!port_table) return -ENOMEM; @@ -1624,21 +1657,7 @@ int tnl_init(void) return 0; } -void tnl_exit(void) +void ovs_tnl_exit(void) { - int i; - - for (i = 0; i < PORT_TABLE_SIZE; i++) { - struct tnl_vport *tnl_vport; - struct hlist_head *hash_head; - struct hlist_node *n; - - hash_head = &port_table[i]; - hlist_for_each_entry(tnl_vport, n, hash_head, hash_node) { - BUG(); - goto out; - } - } -out: kfree(port_table); }