From a7429e55ed9dd4fe486a6579392d9b2e2055257c Mon Sep 17 00:00:00 2001 From: Robert Love Date: Mon, 12 May 2008 17:08:29 -0400 Subject: [PATCH] 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 net: fix tcp_v4_nuke_addr Signed-off-by: Dima Zavin 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 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 Acked-by: Dima Zavin ipv4: disable bottom halves around call to tcp_done(). Signed-off-by: Robert Love Signed-off-by: Colin Cross 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 --- include/linux/sockios.h | 1 + include/net/tcp.h | 2 ++ net/ipv4/af_inet.c | 1 + net/ipv4/devinet.c | 9 ++++++++- net/ipv4/tcp_ipv4.c | 43 +++++++++++++++++++++++++++++++++++++++++ 5 files changed, 55 insertions(+), 1 deletion(-) diff --git a/include/linux/sockios.h b/include/linux/sockios.h index 7997a506ad41..f7ffe36db03c 100644 --- a/include/linux/sockios.h +++ b/include/linux/sockios.h @@ -65,6 +65,7 @@ #define SIOCDIFADDR 0x8936 /* delete PA address */ #define SIOCSIFHWBROADCAST 0x8937 /* set hardware broadcast addr */ #define SIOCGIFCOUNT 0x8938 /* get number of devices */ +#define SIOCKILLADDR 0x8939 /* kill sockets with this local addr */ #define SIOCGIFBR 0x8940 /* Bridging support */ #define SIOCSIFBR 0x8941 /* Set bridging options */ diff --git a/include/net/tcp.h b/include/net/tcp.h index cda30ea354a2..d2b858d1f967 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -1404,6 +1404,8 @@ extern struct sk_buff **tcp4_gro_receive(struct sk_buff **head, extern int tcp_gro_complete(struct sk_buff *skb); extern int tcp4_gro_complete(struct sk_buff *skb); +extern void tcp_v4_nuke_addr(__u32 saddr); + #ifdef CONFIG_PROC_FS extern int tcp4_proc_init(void); extern void tcp4_proc_exit(void); diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index 5683c8f2bfa0..49f024f62963 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -900,6 +900,7 @@ 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 0d4a184af16f..b5aad2bd0445 100644 --- a/net/ipv4/devinet.c +++ b/net/ipv4/devinet.c @@ -59,6 +59,7 @@ #include #include +#include #include #include #include @@ -735,6 +736,7 @@ 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 = -EACCES; if (!capable(CAP_NET_ADMIN)) goto out; @@ -786,7 +788,8 @@ 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) { @@ -912,6 +915,10 @@ 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 = 0; + tcp_v4_nuke_addr(sin->sin_addr.s_addr); + break; } done: rtnl_unlock(); diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index a7d6671e33b8..4186c7613fb4 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -1954,6 +1954,49 @@ void tcp_v4_destroy_sock(struct sock *sk) } EXPORT_SYMBOL(tcp_v4_destroy_sock); +/* + * tcp_v4_nuke_addr - destroy all sockets on the given local address + */ +void tcp_v4_nuke_addr(__u32 saddr) +{ + unsigned int bucket; + + 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 (inet->inet_rcv_saddr != saddr) + continue; + if (sysctl_ip_dynaddr && sk->sk_state == TCP_SYN_SENT) + continue; + if (sock_flag(sk, SOCK_DEAD)) + continue; + + 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); + } +} + #ifdef CONFIG_PROC_FS /* Proc filesystem TCP sock list dumping. */ -- 2.34.1