kernel: Do not allow unicode in nodenames

Verify that unicode isn't used in a node name and logi an info
message if it is. Filter the head part of the name a bit harder
to work with epmd.

2 files changed, 129 insertions, 8 deletions

diff --git a/lib/kernel/src/net_kernel.erl b/lib/kernel/src/net_kernel.erl
index 0a9f9316b0..9921a0adfd 100644
--- a/lib/kernel/src/net_kernel.erl
+++ b/lib/kernel/src/net_kernel.erl
@@ -1262,11 +1262,22 @@ create_name(Name, LongOrShortNames, Try) ->
     {Head,Host1} = create_hostpart(Name, LongOrShortNames),
     case Host1 of
 	{ok,HostPart} ->
-	    {ok,list_to_atom(Head ++ HostPart)};
+            case valid_name_head(Head) of
+                true ->
+                    {ok,list_to_atom(Head ++ HostPart)};
+                false ->
+                    error_logger:info_msg("Invalid node name!\n"
+                                          "Please check your configuration\n"),
+                    {error, badarg}
+            end;
 	{error,long} when Try =:= 1 ->
 	    %% It could be we haven't read domain name from resolv file yet
 	    inet_config:do_load_resolv(os:type(), longnames),
 	    create_name(Name, LongOrShortNames, 0);
+        {error, hostname_not_allowed} ->
+            error_logger:info_msg("Invalid node name!\n"
+                                  "Please check your configuration\n"),
+            {error, badarg};
 	{error,Type} ->
 	    error_logger:info_msg(
 	      lists:concat(["Can\'t set ",
@@ -1279,12 +1290,13 @@ create_name(Name, LongOrShortNames, Try) ->
 create_hostpart(Name, LongOrShortNames) ->
     {Head,Host} = split_node(Name),
     Host1 = case {Host,LongOrShortNames} of
-		{[$@,_|_],longnames} ->
-		    {ok,Host};
+		{[$@,_|_] = Host,longnames} ->
+                    validate_hostname(Host);
 		{[$@,_|_],shortnames} ->
 		    case lists:member($.,Host) of
 			true -> {error,short};
-			_ -> {ok,Host}
+			_ ->
+                            validate_hostname(Host)
 		    end;
 		{_,shortnames} ->
 		    case inet_db:gethostname() of
@@ -1304,6 +1316,24 @@ create_hostpart(Name, LongOrShortNames) ->
 	    end,
     {Head,Host1}.
 
+validate_hostname([$@|HostPart] = Host) ->
+    {ok, MP} = re:compile("^[!-ÿ]*$", [unicode]),
+    case re:run(HostPart, MP) of
+        {match, _} ->
+            {ok, Host};
+        nomatch ->
+            {error, hostname_not_allowed}
+    end.
+
+valid_name_head(Head) ->
+    {ok, MP} = re:compile("^[0-9A-Za-z_\\-]*$", [unicode]),
+        case re:run(Head, MP) of
+            {match, _} ->
+                true;
+            nomatch ->
+                false
+    end.
+
 split_node(Name) ->
     lists:splitwith(fun(C) -> C =/= $@ end, atom_to_list(Name)).
 
diff --git a/lib/kernel/test/erl_distribution_SUITE.erl b/lib/kernel/test/erl_distribution_SUITE.erl
index 09c80a0956..ecfa3d6cdb 100644
--- a/lib/kernel/test/erl_distribution_SUITE.erl
+++ b/lib/kernel/test/erl_distribution_SUITE.erl
@@ -24,7 +24,9 @@
 -export([all/0, suite/0,groups/0,init_per_suite/1, end_per_suite/1, 
 	 init_per_group/2,end_per_group/2]).
 
--export([tick/1, tick_change/1, illegal_nodenames/1, hidden_node/1,
+-export([tick/1, tick_change/1,
+         nodenames/1, hostnames/1,
+         illegal_nodenames/1, hidden_node/1,
 	 setopts/1,
 	 table_waste/1, net_setuptime/1,
 	 inet_dist_options_options/1,
@@ -53,7 +55,6 @@
 
 -export([pinger/1]).
 
-
 -define(DUMMY_NODE,dummy@test01).
 
 %%-----------------------------------------------------------------
@@ -68,8 +69,8 @@ suite() ->
      {timetrap,{minutes,4}}].
 
 all() -> 
-    [tick, tick_change, illegal_nodenames, hidden_node,
-     setopts,
+    [tick, tick_change, nodenames, hostnames, illegal_nodenames,
+     hidden_node, setopts,
      table_waste, net_setuptime, inet_dist_options_options,
      {group, monitor_nodes}].
 
@@ -179,7 +180,97 @@ table_waste(Config) when is_list(Config) ->
     stop_node(N),
     ok.
 
+%% Test that starting nodes with different legal name part works, and that illegal
+%% ones are filtered
+nodenames(Config) when is_list(Config) ->
+    legal("a1@b"),
+    legal("a-1@b"),
+    legal("a_1@b"),
+
+    illegal("cdé@a"),
+    illegal("te欢st@a").
+
+%% Test that starting nodes with different legal host part works, and that illegal
+%% ones are filtered
+hostnames(Config) when is_list(Config) ->
+    Host = gethostname(),
+    legal([$a,$@|atom_to_list(Host)]),
+    legal("1@b1"),
+    legal("b@b1-c"),
+    legal("c@b1_c"),
+    legal("d@b1#c"),
+    legal("f@::1"),
+    legal("g@1:bc3:4e3f:f20:0:1"),
+
+    case file:native_name_encoding() of
+        latin1 -> ignore;
+        _ -> legal("e@b1é")
+    end,
+    long_hostnames(net_kernel:longnames()),
+
+    illegal("h@testالع"),
+    illegal("i@языtest"),
+    illegal("j@te欢st").
+
+long_hostnames(true) ->
+    legal("[email protected]"),
+    legal("[email protected]"),
+    legal("[email protected]_c.d"),
+    legal("[email protected]"),
+    legal("[email protected]");
+long_hostnames(false) ->
+    illegal("[email protected]").
+
+legal(Name) ->
+    case test_node(Name) of
+        started ->
+            ok;
+        not_started ->
+            ct:fail("no ~p node started", [Name])
+    end.
+
+illegal(Name) ->
+    case test_node(Name) of
+        not_started ->
+            ok;
+        started ->
+            ct:fail("~p node started with illegal name", [Name])
+    end.
 
+test_node(Name) ->
+    ProgName = atom_to_list(lib:progname()),
+    Command = ProgName ++ " -noinput " ++ long_or_short() ++ Name ++
+               " -eval \"net_adm:ping('" ++ atom_to_list(node()) ++ "')\"",
+    net_kernel:monitor_nodes(true),
+    BinCommand = unicode:characters_to_binary(Command, utf8),
+    open_port({spawn, BinCommand}, [stream]),
+    Node = list_to_atom(Name),
+    receive
+        {nodeup, Node} ->
+            net_kernel:monitor_nodes(false),
+            slave:stop(Node),
+            started
+    after 5000 ->
+        net_kernel:monitor_nodes(false),
+        not_started
+    end.
+
+long_or_short() ->
+    case net_kernel:longnames() of
+        true -> " -name ";
+        false -> " -sname "
+    end.
+
+% get the localhost's name, depending on the using name policy
+gethostname() ->
+    Hostname = case net_kernel:longnames() of
+       true->
+           net_adm:localhost();
+       _->
+           {ok, Name}=inet:gethostname(),
+           Name
+    end,
+    list_to_atom(Hostname).
 
 %% Test that pinging an illegal nodename does not kill the node.
 illegal_nodenames(Config) when is_list(Config) ->