Update on_load handling during boot

erts/preloaded/src/init.erl

@@ -1,7 +1,7 @@
 %%
 %% %CopyrightBegin%
 %%
-%% Copyright Ericsson AB 1996-2022. All Rights Reserved.
+%% Copyright Ericsson AB 1996-2023. All Rights Reserved.
 %%
 %% Licensed under the Apache License, Version 2.0 (the "License");
 %% you may not use this file except in compliance with the License.
@@ -53,7 +53,7 @@
 	 get_argument/1,script_id/0,script_name/0]).
 
 %% for the on_load functionality; not for general use
--export([run_on_load_handlers/0, run_on_load_handlers/1]).
+-export([run_on_load_handlers/0]).
 
 %% internal exports
 -export([fetch_loaded/0,ensure_loaded/1,make_permanent/2,
@@ -1502,26 +1502,22 @@ archive_extension() ->
 %%%
 
 run_on_load_handlers() ->
-    run_on_load_handlers(all).
-
-run_on_load_handlers(Mods) when is_list(Mods); Mods =:= all ->
     Ref = monitor(process, ?ON_LOAD_HANDLER),
-    catch ?ON_LOAD_HANDLER ! {run_on_load, self(), Ref, Mods},
+    _ = catch ?ON_LOAD_HANDLER ! {run_on_load, Ref},
     receive
 	{'DOWN',Ref,process,_,noproc} ->
 	    %% There is no on_load handler process,
 	    %% probably because init:restart/0 has been
 	    %% called and it is not the first time we
 	    %% pass through here.
 	    ok;
-	{'DOWN',Ref,process,_,on_load_done} ->
+	{'DOWN',Ref,process,_,Ref} ->
+            %% All on_load handlers have run succesfully
 	    ok;
-	{'DOWN',Ref,process,_,Res} ->
+	{'DOWN',Ref,process,_,Reason} ->
 	    %% Failure to run an on_load handler.
 	    %% This is fatal during start-up.
-	    exit(Res);
-        {reply, Ref, on_load_done} ->
-            ok
+	    exit(Reason)
     end.
 
 start_on_load_handler_process() ->
@@ -1537,14 +1533,9 @@ on_load_loop(Mods, Debug0) ->
 	    on_load_loop(Mods, Debug);
 	{loaded,Mod} ->
 	    on_load_loop([Mod|Mods], Debug0);
-	{run_on_load, _, _, all} ->
+	{run_on_load, Ref} ->
 	    run_on_load_handlers(Mods, Debug0),
-	    exit(on_load_done);
-        {run_on_load, From, Ref, ModsToRun} ->
-            [run_on_load_handlers([Mod], Debug0)
-             || Mod <- ModsToRun, lists:member(Mod, Mods)],
-            From ! {reply, Ref, on_load_done},
-            on_load_loop(Mods -- ModsToRun, Debug0)
+	    exit(Ref)
     end.
 
 run_on_load_handlers([M|Ms], Debug) ->

lib/kernel/src/kernel.erl

@@ -1,7 +1,7 @@
 %%
 %% %CopyrightBegin%
 %% 
-%% Copyright Ericsson AB 1996-2022. All Rights Reserved.
+%% Copyright Ericsson AB 1996-2023. All Rights Reserved.
 %% 
 %% Licensed under the Apache License, Version 2.0 (the "License");
 %% you may not use this file except in compliance with the License.
@@ -58,14 +58,19 @@ config_change(Changed, New, Removed) ->
 %%%              | kernel_sup (A)|
 %%%	          ---------------
 %%%                      |
-%%%        -------------------------------
-%%%       |              |                |
-%%%  <std services> -------------   -------------
-%%%   (file,code,  | erl_dist (A)| | safe_sup (1)|
-%%%    rpc, ...)    -------------   -------------
-%%%		          |               |
-%%%                  (net_kernel,  (disk_log, pg,
-%%%          	      auth, ...)     ...)
+%%%        ------------------------------------------
+%%%       |             |          |        |        |
+%%%  <std services>     |    -------------  |  -------------
+%%%   (code, [stderr],  |   | erl_dist (A)| | | safe_sup (1)|
+%%%    ...)             |    -------------  |  -------------
+%%%		        |          |        |        |
+%%%                <on_load>  ([net_sup],   | (disk_log, pg,
+%%%               (transient)  rpc, global, |  ...)
+%%%                            ...)         |
+%%%                                         |
+%%%                                   <more services>
+%%%                                    (file, peer, [user],
+%%%                                     [logger], ...)
 %%%
 %%% The rectangular boxes are supervisors.  All supervisors except
 %%% for kernel_safe_sup terminates the entire erlang node if any of
@@ -118,6 +123,15 @@ init([]) ->
                  type => supervisor,
                  modules => [standard_error]},
 
+    OnLoad = #{id => on_load,
+               start =>
+                   {proc_lib, start_link,
+                    [?MODULE, ?FUNCTION_NAME, [on_load]]},
+               restart => transient,
+               shutdown => 2000,
+               type => worker,
+               modules => [?MODULE]},
+
     User = #{id => user,
              start => {user_sup, start, []},
              restart => temporary,
@@ -148,7 +162,8 @@ init([]) ->
     case init:get_argument(mode) of
         {ok, [["minimal"]|_]} ->
             {ok, {SupFlags,
-                  [Code, File, StdError] ++ Peer ++
+                  [Code, StdError, OnLoad] ++
+                      [File | Peer] ++
                       [User, LoggerSup, Config, RefC, SafeSup]}};
         _ ->
             DistChildren =
@@ -175,11 +190,18 @@ init([]) ->
             CompileServer = start_compile_server(),
 
             {ok, {SupFlags,
-                  [Code, InetDb | DistChildren] ++
-                      [File, SigSrv, StdError] ++ Peer ++
-                      [User, Config, RefC, SafeSup, LoggerSup] ++
+                  [Code, StdError, OnLoad, InetDb | DistChildren] ++
+                      [File, SigSrv | Peer] ++
+                      [User, LoggerSup, Config, RefC, SafeSup] ++
                       Timer ++ CompileServer}}
     end;
+init(on_load) ->
+    %% Run the on_load handlers for all modules that have been
+    %% loaded so far. Running them at this point means that
+    %% on_load handlers can safely call some essential kernel processes,
+    %% in particular call code:priv_dir/1 or code:lib_dir/1.
+    init:run_on_load_handlers(),
+    proc_lib:init_ack({ok, self()});
 init(safe) ->
     SupFlags = #{strategy => one_for_one,
                  intensity => 4,
@@ -189,12 +211,6 @@ init(safe) ->
     DiskLog = start_disk_log(),
     Pg = start_pg(),
 
-    %% Run the on_load handlers for all modules that have been
-    %% loaded so far. Running them at this point means that
-    %% on_load handlers can safely call kernel processes
-    %% (and in particular call code:priv_dir/1 or code:lib_dir/1).
-    init:run_on_load_handlers(),
-
     {ok, {SupFlags, Boot ++ DiskLog ++ Pg}}.
 
 start_distribution() ->

lib/kernel/src/user_drv.erl

@@ -1,7 +1,7 @@
 %%
 %% %CopyrightBegin%
 %% 
-%% Copyright Ericsson AB 1996-2021. All Rights Reserved.
+%% Copyright Ericsson AB 1996-2023. All Rights Reserved.
 %% 
 %% Licensed under the Apache License, Version 2.0 (the "License");
 %% you may not use this file except in compliance with the License.
@@ -139,9 +139,6 @@ callback_mode() -> state_functions.
 -spec init(arguments()) -> gen_statem:init_result(init).
 init(Args) ->
     process_flag(trap_exit, true),
-    %% When running in embedded mode we need to call prim_tty:on_load manually here
-    %% as the automatic call happens after user is started.
-    ok = init:run_on_load_handlers([prim_tty]),
 
     IsTTY = prim_tty:isatty(stdin) =:= true andalso prim_tty:isatty(stdout) =:= true,
     StartShell = maps:get(initial_shell, Args, undefined) =/= noshell,

lib/ssl/src/inet_tls_dist.erl

@@ -1,7 +1,7 @@
 %%
 %% %CopyrightBegin%
 %%
-%% Copyright Ericsson AB 2011-2022. All Rights Reserved.
+%% Copyright Ericsson AB 2011-2023. All Rights Reserved.
 %%
 %% Licensed under the Apache License, Version 2.0 (the "License");
 %% you may not use this file except in compliance with the License.
@@ -291,7 +291,6 @@ spawn_accept({Driver, Listen, Kernel}) ->
 accept_one(Driver, Kernel, Socket) ->
     Opts = setup_verify_client(Socket, get_ssl_options(server)),
     KTLS = proplists:get_value(ktls, Opts, false),
-    wait_for_code_server(),
     case
         ssl:handshake(
           Socket,
@@ -435,36 +434,6 @@ verify_client(PeerCert, valid_peer, {AllowedHosts,PeerIP} = S) ->
     end.
 
 
-wait_for_code_server() ->
-    %% This is an ugly hack.  Upgrading a socket to TLS requires the
-    %% crypto module to be loaded.  Loading the crypto module triggers
-    %% its on_load function, which calls code:priv_dir/1 to find the
-    %% directory where its NIF library is.  However, distribution is
-    %% started earlier than the code server, so the code server is not
-    %% necessarily started yet, and code:priv_dir/1 might fail because
-    %% of that, if we receive an incoming connection on the
-    %% distribution port early enough.
-    %%
-    %% If the on_load function of a module fails, the module is
-    %% unloaded, and the function call that triggered loading it fails
-    %% with 'undef', which is rather confusing.
-    %%
-    %% Thus, the accept process will terminate, and be
-    %% restarted by ssl_dist_sup.  However, it won't have any memory
-    %% of being asked by net_kernel to listen for incoming
-    %% connections.  Hence, the node will believe that it's open for
-    %% distribution, but it actually isn't.
-    %%
-    %% So let's avoid that by waiting for the code server to start.
-    case whereis(code_server) of
-	undefined ->
-	    timer:sleep(10),
-	    wait_for_code_server();
-	Pid when is_pid(Pid) ->
-            init:run_on_load_handlers([crypto,asn1rt_nif]),
-	    ok
-    end.
-
 %% -------------------------------------------------------------------------
 
 accept_connection(AcceptPid, DistCtrl, MyNode, Allowed, SetupTime) ->
@@ -599,8 +568,6 @@ setup_fun(Driver, Kernel, Node, Type, MyNode, LongOrShortNames, SetupTime) ->
 
 -spec do_setup(_,_,_,_,_,_,_) -> no_return().
 do_setup(Driver, Kernel, Node, Type, MyNode, LongOrShortNames, SetupTime) ->
-    wait_for_code_server(),
-
     {Name, Address} = split_node(Driver, Node, LongOrShortNames),
     ErlEpmd = net_kernel:epmd_module(),
     {ARMod, ARFun} = get_address_resolver(ErlEpmd, Driver),