aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorEric Dumazet <eric.dumazet@gmail.com>2010-09-08 01:08:44 -0400
committerDavid S. Miller <davem@davemloft.net>2010-09-09 00:45:01 -0400
commit719f835853a92f6090258114a72ffe41f09155cd (patch)
treea077b05397bf07a096be7f07b50375c5ed918ac2
parentae2688d59b5f861dc70a091d003773975d2ae7fb (diff)
udp: add rehash on connect()
commit 30fff923 introduced in linux-2.6.33 (udp: bind() optimisation) added a secondary hash on UDP, hashed on (local addr, local port). Problem is that following sequence : fd = socket(...) connect(fd, &remote, ...) not only selects remote end point (address and port), but also sets local address, while UDP stack stored in secondary hash table the socket while its local address was INADDR_ANY (or ipv6 equivalent) Sequence is : - autobind() : choose a random local port, insert socket in hash tables [while local address is INADDR_ANY] - connect() : set remote address and port, change local address to IP given by a route lookup. When an incoming UDP frame comes, if more than 10 sockets are found in primary hash table, we switch to secondary table, and fail to find socket because its local address changed. One solution to this problem is to rehash datagram socket if needed. We add a new rehash(struct socket *) method in "struct proto", and implement this method for UDP v4 & v6, using a common helper. This rehashing only takes care of secondary hash table, since primary hash (based on local port only) is not changed. Reported-by: Krzysztof Piotr Oledzki <ole@ans.pl> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Tested-by: Krzysztof Piotr Oledzki <ole@ans.pl> Signed-off-by: David S. Miller <davem@davemloft.net>
-rw-r--r--include/net/sock.h1
-rw-r--r--include/net/udp.h1
-rw-r--r--net/ipv4/datagram.c5
-rw-r--r--net/ipv4/udp.c44
-rw-r--r--net/ipv6/datagram.c7
-rw-r--r--net/ipv6/udp.c10
6 files changed, 66 insertions, 2 deletions
diff --git a/include/net/sock.h b/include/net/sock.h
index ac53bfbdfe16..adab9dc58183 100644
--- a/include/net/sock.h
+++ b/include/net/sock.h
@@ -752,6 +752,7 @@ struct proto {
752 /* Keeping track of sk's, looking them up, and port selection methods. */ 752 /* Keeping track of sk's, looking them up, and port selection methods. */
753 void (*hash)(struct sock *sk); 753 void (*hash)(struct sock *sk);
754 void (*unhash)(struct sock *sk); 754 void (*unhash)(struct sock *sk);
755 void (*rehash)(struct sock *sk);
755 int (*get_port)(struct sock *sk, unsigned short snum); 756 int (*get_port)(struct sock *sk, unsigned short snum);
756 757
757 /* Keeping track of sockets in use */ 758 /* Keeping track of sockets in use */
diff --git a/include/net/udp.h b/include/net/udp.h
index 7abdf305da50..a184d3496b13 100644
--- a/include/net/udp.h
+++ b/include/net/udp.h
@@ -151,6 +151,7 @@ static inline void udp_lib_hash(struct sock *sk)
151} 151}
152 152
153extern void udp_lib_unhash(struct sock *sk); 153extern void udp_lib_unhash(struct sock *sk);
154extern void udp_lib_rehash(struct sock *sk, u16 new_hash);
154 155
155static inline void udp_lib_close(struct sock *sk, long timeout) 156static inline void udp_lib_close(struct sock *sk, long timeout)
156{ 157{
diff --git a/net/ipv4/datagram.c b/net/ipv4/datagram.c
index f0550941df7b..721a8a37b45c 100644
--- a/net/ipv4/datagram.c
+++ b/net/ipv4/datagram.c
@@ -62,8 +62,11 @@ int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
62 } 62 }
63 if (!inet->inet_saddr) 63 if (!inet->inet_saddr)
64 inet->inet_saddr = rt->rt_src; /* Update source address */ 64 inet->inet_saddr = rt->rt_src; /* Update source address */
65 if (!inet->inet_rcv_saddr) 65 if (!inet->inet_rcv_saddr) {
66 inet->inet_rcv_saddr = rt->rt_src; 66 inet->inet_rcv_saddr = rt->rt_src;
67 if (sk->sk_prot->rehash)
68 sk->sk_prot->rehash(sk);
69 }
67 inet->inet_daddr = rt->rt_dst; 70 inet->inet_daddr = rt->rt_dst;
68 inet->inet_dport = usin->sin_port; 71 inet->inet_dport = usin->sin_port;
69 sk->sk_state = TCP_ESTABLISHED; 72 sk->sk_state = TCP_ESTABLISHED;
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index 32e0bef60d0a..fb23c2e63b52 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -1260,6 +1260,49 @@ void udp_lib_unhash(struct sock *sk)
1260} 1260}
1261EXPORT_SYMBOL(udp_lib_unhash); 1261EXPORT_SYMBOL(udp_lib_unhash);
1262 1262
1263/*
1264 * inet_rcv_saddr was changed, we must rehash secondary hash
1265 */
1266void udp_lib_rehash(struct sock *sk, u16 newhash)
1267{
1268 if (sk_hashed(sk)) {
1269 struct udp_table *udptable = sk->sk_prot->h.udp_table;
1270 struct udp_hslot *hslot, *hslot2, *nhslot2;
1271
1272 hslot2 = udp_hashslot2(udptable, udp_sk(sk)->udp_portaddr_hash);
1273 nhslot2 = udp_hashslot2(udptable, newhash);
1274 udp_sk(sk)->udp_portaddr_hash = newhash;
1275 if (hslot2 != nhslot2) {
1276 hslot = udp_hashslot(udptable, sock_net(sk),
1277 udp_sk(sk)->udp_port_hash);
1278 /* we must lock primary chain too */
1279 spin_lock_bh(&hslot->lock);
1280
1281 spin_lock(&hslot2->lock);
1282 hlist_nulls_del_init_rcu(&udp_sk(sk)->udp_portaddr_node);
1283 hslot2->count--;
1284 spin_unlock(&hslot2->lock);
1285
1286 spin_lock(&nhslot2->lock);
1287 hlist_nulls_add_head_rcu(&udp_sk(sk)->udp_portaddr_node,
1288 &nhslot2->head);
1289 nhslot2->count++;
1290 spin_unlock(&nhslot2->lock);
1291
1292 spin_unlock_bh(&hslot->lock);
1293 }
1294 }
1295}
1296EXPORT_SYMBOL(udp_lib_rehash);
1297
1298static void udp_v4_rehash(struct sock *sk)
1299{
1300 u16 new_hash = udp4_portaddr_hash(sock_net(sk),
1301 inet_sk(sk)->inet_rcv_saddr,
1302 inet_sk(sk)->inet_num);
1303 udp_lib_rehash(sk, new_hash);
1304}
1305
1263static int __udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb) 1306static int __udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
1264{ 1307{
1265 int rc; 1308 int rc;
@@ -1843,6 +1886,7 @@ struct proto udp_prot = {
1843 .backlog_rcv = __udp_queue_rcv_skb, 1886 .backlog_rcv = __udp_queue_rcv_skb,
1844 .hash = udp_lib_hash, 1887 .hash = udp_lib_hash,
1845 .unhash = udp_lib_unhash, 1888 .unhash = udp_lib_unhash,
1889 .rehash = udp_v4_rehash,
1846 .get_port = udp_v4_get_port, 1890 .get_port = udp_v4_get_port,
1847 .memory_allocated = &udp_memory_allocated, 1891 .memory_allocated = &udp_memory_allocated,
1848 .sysctl_mem = sysctl_udp_mem, 1892 .sysctl_mem = sysctl_udp_mem,
diff --git a/net/ipv6/datagram.c b/net/ipv6/datagram.c
index 7d929a22cbc2..ef371aa01ac5 100644
--- a/net/ipv6/datagram.c
+++ b/net/ipv6/datagram.c
@@ -105,9 +105,12 @@ ipv4_connected:
105 if (ipv6_addr_any(&np->saddr)) 105 if (ipv6_addr_any(&np->saddr))
106 ipv6_addr_set_v4mapped(inet->inet_saddr, &np->saddr); 106 ipv6_addr_set_v4mapped(inet->inet_saddr, &np->saddr);
107 107
108 if (ipv6_addr_any(&np->rcv_saddr)) 108 if (ipv6_addr_any(&np->rcv_saddr)) {
109 ipv6_addr_set_v4mapped(inet->inet_rcv_saddr, 109 ipv6_addr_set_v4mapped(inet->inet_rcv_saddr,
110 &np->rcv_saddr); 110 &np->rcv_saddr);
111 if (sk->sk_prot->rehash)
112 sk->sk_prot->rehash(sk);
113 }
111 114
112 goto out; 115 goto out;
113 } 116 }
@@ -181,6 +184,8 @@ ipv4_connected:
181 if (ipv6_addr_any(&np->rcv_saddr)) { 184 if (ipv6_addr_any(&np->rcv_saddr)) {
182 ipv6_addr_copy(&np->rcv_saddr, &fl.fl6_src); 185 ipv6_addr_copy(&np->rcv_saddr, &fl.fl6_src);
183 inet->inet_rcv_saddr = LOOPBACK4_IPV6; 186 inet->inet_rcv_saddr = LOOPBACK4_IPV6;
187 if (sk->sk_prot->rehash)
188 sk->sk_prot->rehash(sk);
184 } 189 }
185 190
186 ip6_dst_store(sk, dst, 191 ip6_dst_store(sk, dst,
diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
index 1dd1affdead2..5acb3560ff15 100644
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -111,6 +111,15 @@ int udp_v6_get_port(struct sock *sk, unsigned short snum)
111 return udp_lib_get_port(sk, snum, ipv6_rcv_saddr_equal, hash2_nulladdr); 111 return udp_lib_get_port(sk, snum, ipv6_rcv_saddr_equal, hash2_nulladdr);
112} 112}
113 113
114static void udp_v6_rehash(struct sock *sk)
115{
116 u16 new_hash = udp6_portaddr_hash(sock_net(sk),
117 &inet6_sk(sk)->rcv_saddr,
118 inet_sk(sk)->inet_num);
119
120 udp_lib_rehash(sk, new_hash);
121}
122
114static inline int compute_score(struct sock *sk, struct net *net, 123static inline int compute_score(struct sock *sk, struct net *net,
115 unsigned short hnum, 124 unsigned short hnum,
116 struct in6_addr *saddr, __be16 sport, 125 struct in6_addr *saddr, __be16 sport,
@@ -1447,6 +1456,7 @@ struct proto udpv6_prot = {
1447 .backlog_rcv = udpv6_queue_rcv_skb, 1456 .backlog_rcv = udpv6_queue_rcv_skb,
1448 .hash = udp_lib_hash, 1457 .hash = udp_lib_hash,
1449 .unhash = udp_lib_unhash, 1458 .unhash = udp_lib_unhash,
1459 .rehash = udp_v6_rehash,
1450 .get_port = udp_v6_get_port, 1460 .get_port = udp_v6_get_port,
1451 .memory_allocated = &udp_memory_allocated, 1461 .memory_allocated = &udp_memory_allocated,
1452 .sysctl_mem = sysctl_udp_mem, 1462 .sysctl_mem = sysctl_udp_mem,