diff options
author | NeilBrown <neilb@suse.de> | 2013-10-31 01:14:36 -0400 |
---|---|---|
committer | Trond Myklebust <Trond.Myklebust@netapp.com> | 2013-10-31 09:14:50 -0400 |
commit | 93dc41bdc5c853916610576c6b48a1704959c70d (patch) | |
tree | 893c971fc8538c8872938c56a442be0abe7db437 /net/sunrpc | |
parent | 09c3e54635c85b3da44d3bc156619c1f1af3bb43 (diff) |
SUNRPC: close a rare race in xs_tcp_setup_socket.
We have one report of a crash in xs_tcp_setup_socket.
The call path to the crash is:
xs_tcp_setup_socket -> inet_stream_connect -> lock_sock_nested.
The 'sock' passed to that last function is NULL.
The only way I can see this happening is a concurrent call to
xs_close:
xs_close -> xs_reset_transport -> sock_release -> inet_release
inet_release sets:
sock->sk = NULL;
inet_stream_connect calls
lock_sock(sock->sk);
which gets NULL.
All calls to xs_close are protected by XPRT_LOCKED as are most
activations of the workqueue which runs xs_tcp_setup_socket.
The exception is xs_tcp_schedule_linger_timeout.
So presumably the timeout queued by the later fires exactly when some
other code runs xs_close().
To protect against this we can move the cancel_delayed_work_sync()
call from xs_destory() to xs_close().
As xs_close is never called from the worker scheduled on
->connect_worker, this can never deadlock.
Signed-off-by: NeilBrown <neilb@suse.de>
[Trond: Make it safe to call cancel_delayed_work_sync() on AF_LOCAL sockets]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Diffstat (limited to 'net/sunrpc')
-rw-r--r-- | net/sunrpc/xprtsock.c | 13 |
1 files changed, 9 insertions, 4 deletions
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index 9deed17fd3e4..a4709bbf8e5e 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c | |||
@@ -835,6 +835,8 @@ static void xs_close(struct rpc_xprt *xprt) | |||
835 | 835 | ||
836 | dprintk("RPC: xs_close xprt %p\n", xprt); | 836 | dprintk("RPC: xs_close xprt %p\n", xprt); |
837 | 837 | ||
838 | cancel_delayed_work_sync(&transport->connect_worker); | ||
839 | |||
838 | xs_reset_transport(transport); | 840 | xs_reset_transport(transport); |
839 | xprt->reestablish_timeout = 0; | 841 | xprt->reestablish_timeout = 0; |
840 | 842 | ||
@@ -869,12 +871,8 @@ static void xs_local_destroy(struct rpc_xprt *xprt) | |||
869 | */ | 871 | */ |
870 | static void xs_destroy(struct rpc_xprt *xprt) | 872 | static void xs_destroy(struct rpc_xprt *xprt) |
871 | { | 873 | { |
872 | struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt); | ||
873 | |||
874 | dprintk("RPC: xs_destroy xprt %p\n", xprt); | 874 | dprintk("RPC: xs_destroy xprt %p\n", xprt); |
875 | 875 | ||
876 | cancel_delayed_work_sync(&transport->connect_worker); | ||
877 | |||
878 | xs_local_destroy(xprt); | 876 | xs_local_destroy(xprt); |
879 | } | 877 | } |
880 | 878 | ||
@@ -1817,6 +1815,10 @@ static inline void xs_reclassify_socket(int family, struct socket *sock) | |||
1817 | } | 1815 | } |
1818 | #endif | 1816 | #endif |
1819 | 1817 | ||
1818 | static void xs_dummy_setup_socket(struct work_struct *work) | ||
1819 | { | ||
1820 | } | ||
1821 | |||
1820 | static struct socket *xs_create_sock(struct rpc_xprt *xprt, | 1822 | static struct socket *xs_create_sock(struct rpc_xprt *xprt, |
1821 | struct sock_xprt *transport, int family, int type, int protocol) | 1823 | struct sock_xprt *transport, int family, int type, int protocol) |
1822 | { | 1824 | { |
@@ -2668,6 +2670,9 @@ static struct rpc_xprt *xs_setup_local(struct xprt_create *args) | |||
2668 | xprt->ops = &xs_local_ops; | 2670 | xprt->ops = &xs_local_ops; |
2669 | xprt->timeout = &xs_local_default_timeout; | 2671 | xprt->timeout = &xs_local_default_timeout; |
2670 | 2672 | ||
2673 | INIT_DELAYED_WORK(&transport->connect_worker, | ||
2674 | xs_dummy_setup_socket); | ||
2675 | |||
2671 | switch (sun->sun_family) { | 2676 | switch (sun->sun_family) { |
2672 | case AF_LOCAL: | 2677 | case AF_LOCAL: |
2673 | if (sun->sun_path[0] != '/') { | 2678 | if (sun->sun_path[0] != '/') { |