We saw a report of a NULL dereference in xprt_autoclose:
	https://bugzilla.redhat.com/show_bug.cgi?id=611938
This appears to be the result of an xprt's task_cleanup running after
the xprt is destroyed.  Nothing in the current code appears to prevent
that.
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
        rpc_destroy_wait_queue(&xprt->sending);
        rpc_destroy_wait_queue(&xprt->resend);
        rpc_destroy_wait_queue(&xprt->backlog);
+       cancel_work_sync(&xprt->task_cleanup);
        /*
         * Tear down transport state and free the rpc_xprt
         */