net: socket ioctl to reset connections matching local address

Introduce a new socket ioctl, SIOCKILLADDR, that nukes all sockets
bound to the same local address. This is useful in situations with
dynamic IPs, to kill stuck connections.

Signed-off-by: Brian Swetland <swetland@google.com>

net: fix tcp_v4_nuke_addr

Signed-off-by: Dima Zavin <dima@android.com>

net: ipv4: Fix a spinlock recursion bug in tcp_v4_nuke.

We can't hold the lock while calling to tcp_done(), so we drop
it before calling. We then have to start at the top of the chain again.

Signed-off-by: Dima Zavin <dima@android.com>

net: ipv4: Fix race in tcp_v4_nuke_addr().

To fix a recursive deadlock in 2.6.29, we stopped holding the hash table lock
across tcp_done() calls. This fixed the deadlock, but introduced a race where
the socket could die or change state.

Fix: Before unlocking the hash table, we grab a reference to the socket. We
can then unlock the hash table without risk of the socket going away. We then
lock the socket, which is safe because it is pinned. We can then call
tcp_done() without recursive deadlock and without race. Upon return, we unlock
the socket and then unpin it, killing it.

Change-Id: Idcdae072b48238b01bdbc8823b60310f1976e045
Signed-off-by: Robert Love <rlove@google.com>
Acked-by: Dima Zavin <dima@android.com>

ipv4: disable bottom halves around call to tcp_done().

Signed-off-by: Robert Love <rlove@google.com>
Signed-off-by: Colin Cross <ccross@android.com>

ipv4: Move sk_error_report inside bh_lock_sock in tcp_v4_nuke_addr

When sk_error_report is called, it wakes up the user-space thread, which then
calls tcp_close.  When the tcp_close is interrupted by the tcp_v4_nuke_addr
ioctl thread running tcp_done, it leaks 392 bytes and triggers a WARN_ON.

This patch moves the call to sk_error_report inside the bh_lock_sock, which
matches the locking used in tcp_v4_err.

Signed-off-by: Colin Cross <ccross@android.com>

4 files changed, 132 insertions, 1 deletions

diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 0b630d2406b..ad3c54bb090 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -936,6 +936,7 @@ int inet_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
 	case SIOCSIFPFLAGS:
 	case SIOCGIFPFLAGS:
 	case SIOCSIFFLAGS:
+	case SIOCKILLADDR:
 		err = devinet_ioctl(net, cmd, (void __user *)arg);
 		break;
 	default:
diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c
index dfc39d4d48b..b326e6754e2 100644
--- a/net/ipv4/devinet.c
+++ b/net/ipv4/devinet.c
@@ -59,6 +59,7 @@
 
 #include <net/arp.h>
 #include <net/ip.h>
+#include <net/tcp.h>
 #include <net/route.h>
 #include <net/ip_fib.h>
 #include <net/rtnetlink.h>
@@ -916,6 +917,7 @@ int devinet_ioctl(struct net *net, unsigned int cmd, void __user *arg)
 	case SIOCSIFBRDADDR:	/* Set the broadcast address */
 	case SIOCSIFDSTADDR:	/* Set the destination address */
 	case SIOCSIFNETMASK: 	/* Set the netmask for the interface */
+	case SIOCKILLADDR:	/* Nuke all sockets on this address */
 		ret = -EPERM;
 		if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
 			goto out;
@@ -967,7 +969,8 @@ int devinet_ioctl(struct net *net, unsigned int cmd, void __user *arg)
 	}
 
 	ret = -EADDRNOTAVAIL;
-	if (!ifa && cmd != SIOCSIFADDR && cmd != SIOCSIFFLAGS)
+	if (!ifa && cmd != SIOCSIFADDR && cmd != SIOCSIFFLAGS
+	    && cmd != SIOCKILLADDR)
 		goto done;
 
 	switch (cmd) {
@@ -1094,6 +1097,9 @@ int devinet_ioctl(struct net *net, unsigned int cmd, void __user *arg)
 			inet_insert_ifa(ifa);
 		}
 		break;
+	case SIOCKILLADDR:	/* Nuke all connections on this address */
+		ret = tcp_nuke_addr(net, (struct sockaddr *) sin);
+		break;
 	}
 done:
 	rtnl_unlock();
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index ab450c099aa..0ce860b93c1 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -274,6 +274,9 @@
 #include <net/tcp.h>
 #include <net/xfrm.h>
 #include <net/ip.h>
+#include <net/ip6_route.h>
+#include <net/ipv6.h>
+#include <net/transp_v6.h>
 #include <net/netdma.h>
 #include <net/sock.h>
 
@@ -3448,3 +3451,107 @@ void __init tcp_init(void)
 
 	tcp_tasklet_init();
 }
+
+static int tcp_is_local(struct net *net, __be32 addr) {
+	struct rtable *rt;
+	struct flowi4 fl4 = { .daddr = addr };
+	rt = ip_route_output_key(net, &fl4);
+	if (IS_ERR_OR_NULL(rt))
+		return 0;
+	return rt->dst.dev && (rt->dst.dev->flags & IFF_LOOPBACK);
+}
+
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+static int tcp_is_local6(struct net *net, struct in6_addr *addr) {
+	struct rt6_info *rt6 = rt6_lookup(net, addr, addr, 0, 0);
+	return rt6 && rt6->dst.dev && (rt6->dst.dev->flags & IFF_LOOPBACK);
+}
+#endif
+
+/*
+ * tcp_nuke_addr - destroy all sockets on the given local address
+ * if local address is the unspecified address (0.0.0.0 or ::), destroy all
+ * sockets with local addresses that are not configured.
+ */
+int tcp_nuke_addr(struct net *net, struct sockaddr *addr)
+{
+	int family = addr->sa_family;
+	unsigned int bucket;
+
+	struct in_addr *in;
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+	struct in6_addr *in6;
+#endif
+	if (family == AF_INET) {
+		in = &((struct sockaddr_in *)addr)->sin_addr;
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+	} else if (family == AF_INET6) {
+		in6 = &((struct sockaddr_in6 *)addr)->sin6_addr;
+#endif
+	} else {
+		return -EAFNOSUPPORT;
+	}
+
+	for (bucket = 0; bucket < tcp_hashinfo.ehash_mask; bucket++) {
+		struct hlist_nulls_node *node;
+		struct sock *sk;
+		spinlock_t *lock = inet_ehash_lockp(&tcp_hashinfo, bucket);
+
+restart:
+		spin_lock_bh(lock);
+		sk_nulls_for_each(sk, node, &tcp_hashinfo.ehash[bucket].chain) {
+			struct inet_sock *inet = inet_sk(sk);
+
+			if (sysctl_ip_dynaddr && sk->sk_state == TCP_SYN_SENT)
+				continue;
+			if (sock_flag(sk, SOCK_DEAD))
+				continue;
+
+			if (family == AF_INET) {
+				__be32 s4 = inet->inet_rcv_saddr;
+				if (s4 == LOOPBACK4_IPV6)
+					continue;
+
+				if (in->s_addr != s4 &&
+				    !(in->s_addr == INADDR_ANY &&
+				      !tcp_is_local(net, s4)))
+					continue;
+			}
+
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+			if (family == AF_INET6) {
+				struct in6_addr *s6;
+				if (!inet->pinet6)
+					continue;
+
+				s6 = &inet->pinet6->rcv_saddr;
+				if (ipv6_addr_type(s6) == IPV6_ADDR_MAPPED)
+					continue;
+
+				if (!ipv6_addr_equal(in6, s6) &&
+				    !(ipv6_addr_equal(in6, &in6addr_any) &&
+				      !tcp_is_local6(net, s6)))
+				continue;
+			}
+#endif
+
+			sock_hold(sk);
+			spin_unlock_bh(lock);
+
+			local_bh_disable();
+			bh_lock_sock(sk);
+			sk->sk_err = ETIMEDOUT;
+			sk->sk_error_report(sk);
+
+			tcp_done(sk);
+			bh_unlock_sock(sk);
+			local_bh_enable();
+			sock_put(sk);
+
+			goto restart;
+		}
+		spin_unlock_bh(lock);
+	}
+
+	return 0;
+}
diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c
index 4ddbb8086b3..3257ab81a73 100644
--- a/net/ipv6/af_inet6.c
+++ b/net/ipv6/af_inet6.c
@@ -507,6 +507,21 @@ int inet6_getname(struct socket *sock, struct sockaddr *uaddr,
 }
 EXPORT_SYMBOL(inet6_getname);
 
+int inet6_killaddr_ioctl(struct net *net, void __user *arg) {
+	struct in6_ifreq ireq;
+	struct sockaddr_in6 sin6;
+
+	if (!capable(CAP_NET_ADMIN))
+		return -EACCES;
+
+	if (copy_from_user(&ireq, arg, sizeof(struct in6_ifreq)))
+		return -EFAULT;
+
+	sin6.sin6_family = AF_INET6;
+	sin6.sin6_addr = ireq.ifr6_addr;
+	return tcp_nuke_addr(net, (struct sockaddr *) &sin6);
+}
+
 int inet6_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
 {
 	struct sock *sk = sock->sk;
@@ -530,6 +545,8 @@ int inet6_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
 		return addrconf_del_ifaddr(net, (void __user *) arg);
 	case SIOCSIFDSTADDR:
 		return addrconf_set_dstaddr(net, (void __user *) arg);
+	case SIOCKILLADDR:
+		return inet6_killaddr_ioctl(net, (void __user *) arg);
 	default:
 		if (!sk->sk_prot->ioctl)
 			return -ENOIOCTLCMD;