SUNRPC: add WARN_ON_ONCE for potential deadlock

westonandrosadamson · Trond Myklebust · commit 168e4b39d1af · 2012-11-04T14:43:38.000-05:00
rpc_shutdown_client should never be called from a workqueue context.
If it is, it could deadlock looping forever trying to kill tasks that are
assigned to the same kworker thread (and will never run rpc_exit_task).

Signed-off-by: Weston Andros Adamson &lt;dros@netapp.com&gt;
Signed-off-by: Trond Myklebust &lt;Trond.Myklebust@netapp.com&gt;
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
@@ -607,6 +607,13 @@ EXPORT_SYMBOL_GPL(rpc_killall_tasks);
  */
 void rpc_shutdown_client(struct rpc_clnt *clnt)
 {
+	/*
+	 * To avoid deadlock, never call rpc_shutdown_client from a
+	 * workqueue context!
+	 */
+	WARN_ON_ONCE(current->flags & PF_WQ_WORKER);
+	might_sleep();
+
 	dprintk_rcu("RPC:       shutting down %s client for %s\n",
 			clnt->cl_protname,
 			rcu_dereference(clnt->cl_xprt)->servername);
