summaryrefslogtreecommitdiffstats
path: root/net/ipv6
diff options
context:
space:
mode:
authorCraig Gallek <kraig@google.com>2016-01-04 23:41:46 +0100
committerDavid S. Miller <davem@davemloft.net>2016-01-05 04:49:58 +0100
commite32ea7e747271a0abcd37e265005e97cc81d9df5 (patch)
treef3176589179db2e0df7a306bb3564ef6c6fddc17 /net/ipv6
parentsoreuseport: define reuseport groups (diff)
downloadlinux-e32ea7e747271a0abcd37e265005e97cc81d9df5.tar.xz
linux-e32ea7e747271a0abcd37e265005e97cc81d9df5.zip
soreuseport: fast reuseport UDP socket selection
Include a struct sock_reuseport instance when a UDP socket binds to a specific address for the first time with the reuseport flag set. When selecting a socket for an incoming UDP packet, use the information available in sock_reuseport if present. This required adding an additional field to the UDP source address equality function to differentiate between exact and wildcard matches. The original use case allowed wildcard matches when checking for existing port uses during bind. The new use case of adding a socket to a reuseport group requires exact address matching. Performance test (using a machine with 2 CPU sockets and a total of 48 cores): Create reuseport groups of varying size. Use one socket from this group per user thread (pinning each thread to a different core) calling recvmmsg in a tight loop. Record number of messages received per second while saturating a 10G link. 10 sockets: 18% increase (~2.8M -> 3.3M pkts/s) 20 sockets: 14% increase (~2.9M -> 3.3M pkts/s) 40 sockets: 13% increase (~3.0M -> 3.4M pkts/s) This work is based off a similar implementation written by Ying Cai <ycai@google.com> for implementing policy-based reuseport selection. Signed-off-by: Craig Gallek <kraig@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/ipv6')
-rw-r--r--net/ipv6/inet6_connection_sock.c4
-rw-r--r--net/ipv6/udp.c48
2 files changed, 41 insertions, 11 deletions
diff --git a/net/ipv6/inet6_connection_sock.c b/net/ipv6/inet6_connection_sock.c
index a7ca2cde2ecb..36c3f0155010 100644
--- a/net/ipv6/inet6_connection_sock.c
+++ b/net/ipv6/inet6_connection_sock.c
@@ -51,12 +51,12 @@ int inet6_csk_bind_conflict(const struct sock *sk,
(sk2->sk_state != TCP_TIME_WAIT &&
!uid_eq(uid,
sock_i_uid((struct sock *)sk2))))) {
- if (ipv6_rcv_saddr_equal(sk, sk2))
+ if (ipv6_rcv_saddr_equal(sk, sk2, true))
break;
}
if (!relax && reuse && sk2->sk_reuse &&
sk2->sk_state != TCP_LISTEN &&
- ipv6_rcv_saddr_equal(sk, sk2))
+ ipv6_rcv_saddr_equal(sk, sk2, true))
break;
}
}
diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
index 00775ee27d86..6204b8992de4 100644
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -47,6 +47,7 @@
#include <net/xfrm.h>
#include <net/inet6_hashtables.h>
#include <net/busy_poll.h>
+#include <net/sock_reuseport.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
@@ -76,7 +77,14 @@ static u32 udp6_ehashfn(const struct net *net,
udp_ipv6_hash_secret + net_hash_mix(net));
}
-int ipv6_rcv_saddr_equal(const struct sock *sk, const struct sock *sk2)
+/* match_wildcard == true: IPV6_ADDR_ANY equals to any IPv6 addresses if IPv6
+ * only, and any IPv4 addresses if not IPv6 only
+ * match_wildcard == false: addresses must be exactly the same, i.e.
+ * IPV6_ADDR_ANY only equals to IPV6_ADDR_ANY,
+ * and 0.0.0.0 equals to 0.0.0.0 only
+ */
+int ipv6_rcv_saddr_equal(const struct sock *sk, const struct sock *sk2,
+ bool match_wildcard)
{
const struct in6_addr *sk2_rcv_saddr6 = inet6_rcv_saddr(sk2);
int sk2_ipv6only = inet_v6_ipv6only(sk2);
@@ -84,16 +92,24 @@ int ipv6_rcv_saddr_equal(const struct sock *sk, const struct sock *sk2)
int addr_type2 = sk2_rcv_saddr6 ? ipv6_addr_type(sk2_rcv_saddr6) : IPV6_ADDR_MAPPED;
/* if both are mapped, treat as IPv4 */
- if (addr_type == IPV6_ADDR_MAPPED && addr_type2 == IPV6_ADDR_MAPPED)
- return (!sk2_ipv6only &&
- (!sk->sk_rcv_saddr || !sk2->sk_rcv_saddr ||
- sk->sk_rcv_saddr == sk2->sk_rcv_saddr));
+ if (addr_type == IPV6_ADDR_MAPPED && addr_type2 == IPV6_ADDR_MAPPED) {
+ if (!sk2_ipv6only) {
+ if (sk->sk_rcv_saddr == sk2->sk_rcv_saddr)
+ return 1;
+ if (!sk->sk_rcv_saddr || !sk2->sk_rcv_saddr)
+ return match_wildcard;
+ }
+ return 0;
+ }
+
+ if (addr_type == IPV6_ADDR_ANY && addr_type2 == IPV6_ADDR_ANY)
+ return 1;
- if (addr_type2 == IPV6_ADDR_ANY &&
+ if (addr_type2 == IPV6_ADDR_ANY && match_wildcard &&
!(sk2_ipv6only && addr_type == IPV6_ADDR_MAPPED))
return 1;
- if (addr_type == IPV6_ADDR_ANY &&
+ if (addr_type == IPV6_ADDR_ANY && match_wildcard &&
!(ipv6_only_sock(sk) && addr_type2 == IPV6_ADDR_MAPPED))
return 1;
@@ -253,8 +269,14 @@ begin:
badness = score;
reuseport = sk->sk_reuseport;
if (reuseport) {
+ struct sock *sk2;
hash = udp6_ehashfn(net, daddr, hnum,
saddr, sport);
+ sk2 = reuseport_select_sock(sk, hash);
+ if (sk2) {
+ result = sk2;
+ goto found;
+ }
matches = 1;
}
} else if (score == badness && reuseport) {
@@ -273,6 +295,7 @@ begin:
goto begin;
if (result) {
+found:
if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
result = NULL;
else if (unlikely(compute_score2(result, net, saddr, sport,
@@ -332,8 +355,14 @@ begin:
badness = score;
reuseport = sk->sk_reuseport;
if (reuseport) {
+ struct sock *sk2;
hash = udp6_ehashfn(net, daddr, hnum,
saddr, sport);
+ sk2 = reuseport_select_sock(sk, hash);
+ if (sk2) {
+ result = sk2;
+ goto found;
+ }
matches = 1;
}
} else if (score == badness && reuseport) {
@@ -352,6 +381,7 @@ begin:
goto begin;
if (result) {
+found:
if (unlikely(!atomic_inc_not_zero_hint(&result->sk_refcnt, 2)))
result = NULL;
else if (unlikely(compute_score(result, net, hnum, saddr, sport,
@@ -549,8 +579,8 @@ void __udp6_lib_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
int err;
struct net *net = dev_net(skb->dev);
- sk = __udp6_lib_lookup(net, daddr, uh->dest,
- saddr, uh->source, inet6_iif(skb), udptable);
+ sk = __udp6_lib_lookup(net, daddr, uh->dest, saddr, uh->source,
+ inet6_iif(skb), udptable);
if (!sk) {
ICMP6_INC_STATS_BH(net, __in6_dev_get(skb->dev),
ICMP6_MIB_INERRORS);