From d3e5383f3647ed55f2f91355083f3ad55d358344 Mon Sep 17 00:00:00 2001
From: Mikael Pettersson <mikpelinux@gmail.com>
Date: Sat, 7 Sep 2019 17:40:39 +0200
Subject: [PATCH] as: support local labels

---
 erlang/apps/as/priv/test3.s  | 28 ++++++++++++++++++++++++
 erlang/apps/as/src/input.erl | 41 ++++++++++++++++++++++++++++++++----
 erlang/apps/as/src/parse.erl | 25 ++++++++++++++++++++--
 erlang/apps/as/src/scan.erl  |  3 ++-
 erlang/apps/as/src/token.erl |  4 +++-
 erlang/apps/as/src/token.hrl |  6 ++++--
 erlang/apps/as/src/tunit.erl | 11 ++++++++++
 erlang/apps/as/src/tunit.hrl |  7 +++++-
 8 files changed, 114 insertions(+), 11 deletions(-)
 create mode 100644 erlang/apps/as/priv/test3.s

diff --git a/erlang/apps/as/priv/test3.s b/erlang/apps/as/priv/test3.s
new file mode 100644
index 0000000..61f728d
--- /dev/null
+++ b/erlang/apps/as/priv/test3.s
@@ -0,0 +1,28 @@
+/*
+ * test3.s
+ * Copyright (C) 2019  Mikael Pettersson
+ *
+ * This file is part of pdp10-tools.
+ *
+ * pdp10-tools is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * pdp10-tools is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with pdp10-tools.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+	.file	"test3.s"
+	.text
+	.globl	foo
+	.type	foo,@function
+foo:	jumpa	1f
+0:	popj	017,
+1:	jumpa	0b
+	.size	foo,.-foo
diff --git a/erlang/apps/as/src/input.erl b/erlang/apps/as/src/input.erl
index a950666..51ce550 100644
--- a/erlang/apps/as/src/input.erl
+++ b/erlang/apps/as/src/input.erl
@@ -80,6 +80,7 @@ interpret(ScanState, Tunit, Stmt) ->
     #s_dot_text{} -> dot_text(ScanState, Tunit, Stmt);
     #s_dot_type{} -> dot_type(ScanState, Tunit, Stmt);
     #s_label{} -> label(ScanState, Tunit, Stmt);
+    #s_local_label{} -> local_label(ScanState, Tunit, Stmt);
     #s_insn{} -> insn(ScanState, Tunit, Stmt)
   end.
 
@@ -175,29 +176,61 @@ dot_type(ScanState, Tunit, #s_dot_type{name = Name}) ->
 
 label(ScanState, Tunit, #s_label{name = Name}) ->
   case tunit:get_symbol(Tunit, Name) of
-    #symbol{section = false, st_value = false} = Symbol -> label2(Tunit, Symbol);
+    #symbol{section = false, st_value = false} = Symbol -> define_label(Tunit, Symbol);
     #symbol{} -> fmterr(ScanState, "label ~s already defined", [Name]);
-    false -> label2(Tunit, #symbol{name = Name, st_size = false, st_info = 0})
+    false -> define_new_label(Tunit, Name)
   end.
 
-label2(Tunit, Symbol) ->
+define_new_label(Tunit, Name) ->
+  define_label(Tunit, #symbol{name = Name, st_size = false, st_info = 0}).
+
+define_label(Tunit, Symbol) ->
   #tunit{cursect = Cursect} = Tunit,
   #section{dot = Dot} = tunit:get_section(Tunit, Cursect),
   {ok, tunit:put_symbol(Tunit, Symbol#symbol{section = Cursect, st_value = Dot})}.
 
+local_label(_ScanState, Tunit, #s_local_label{number = Number}) ->
+  Serial = local_label_serial(Tunit, Number) + 1,
+  Name = local_label_name(Number, Serial),
+  define_new_label(tunit:put_local_label(Tunit, Number, Serial), Name).
+
+local_label_serial(Tunit, Number) ->
+  case tunit:get_local_label(Tunit, Number) of
+    false -> 0;
+    Serial -> Serial
+  end.
+
+local_label_name(Number, Serial) ->
+  lists:flatten(io_lib:format(".L~.10b\^B~.10b", [Number, Serial])).
+
 insn(ScanState, Tunit, #s_insn{} = Stmt) ->
   #tunit{cursect = Cursect} = Tunit,
   #section{data = {stmts, Stmts}, dot = Dot} = Section = tunit:get_section(Tunit, Cursect),
   case Dot rem 4 of % FIXME: target-specific
     0 ->
+      NewStmt = insn_fixup(Tunit, Stmt),
       NewSection =
-        Section#section{ data = {stmts, [Stmt | Stmts]}
+        Section#section{ data = {stmts, [NewStmt | Stmts]}
                        , dot = Dot + 4 % FIXME: target-specific
                        },
       {ok, tunit:put_section(Tunit, NewSection)};
     _ -> fmterr(ScanState, "misaligned address for instruction", [])
   end.
 
+insn_fixup(Tunit, Insn) ->
+  case Insn#s_insn.address of
+    #e_local_label{number = Number, direction = Direction} ->
+      LabelSerial = local_label_serial(Tunit, Number),
+      ReferenceSerial =
+        case Direction of
+          $b -> LabelSerial;
+          $f -> LabelSerial + 1
+        end,
+      Name = local_label_name(Number, ReferenceSerial),
+      Insn#s_insn{address = #e_symbol{name = Name}};
+    _ -> Insn
+  end.
+
 %% Initialization --------------------------------------------------------------
 
 tunit_init() ->
diff --git a/erlang/apps/as/src/parse.erl b/erlang/apps/as/src/parse.erl
index 1969291..cdd5779 100644
--- a/erlang/apps/as/src/parse.erl
+++ b/erlang/apps/as/src/parse.erl
@@ -39,6 +39,7 @@ stmt(ScanState) ->
     {ok, ?T_DOT_TEXT} -> dot_text(ScanState);
     {ok, ?T_DOT_TYPE} -> dot_type(ScanState);
     {ok, {?T_SYMBOL, Name}} -> stmt_after_symbol(ScanState, Name);
+    {ok, {?T_UINTEGER, UInt}} -> stmt_after_uinteger(ScanState, UInt);
     {ok, ?T_NEWLINE} -> stmt(ScanState);
     {ok, ?T_EOF} -> eof;
     ScanRes -> badtok(ScanState, "expected directive, label, or instruction", ScanRes)
@@ -48,7 +49,7 @@ stmt(ScanState) ->
 %%
 %% Recognize:
 %%
-%% <label> ::= <symbol> ":"
+%% <label> ::= <symbol> ":" | <uinteger> ":"
 %%
 %% <insn> ::= <symbol> (<accumulator> ",")? <address> <newline>
 %%
@@ -56,7 +57,7 @@ stmt(ScanState) ->
 %%
 %% <address> ::= "@"? <displacement>? <index>?
 %%
-%% <displacement> ::= <uinteger> | <symbol>
+%% <displacement> ::= <uinteger> | <symbol> | <uinteger>[bf]
 %%
 %% <index> ::= "(" <accumulator> ")"
 %%
@@ -94,6 +95,15 @@ stmt_after_symbol(ScanState, Name) ->
     {ok, ?T_NEWLINE} -> make_insn(ScanState, Name, false, false, false, false);
     {ok, {?T_UINTEGER, UInt}} -> insn_uint(ScanState, Name, UInt);
     {ok, {?T_SYMBOL, Symbol}} -> insn_symbol(ScanState, Name, Symbol);
+    {ok, {?T_LOCAL_LABEL, Number, Direction}} ->
+      insn_local_label(ScanState, Name, Number, Direction);
+    ScanRes -> badtok(ScanState, "junk after symbol", ScanRes)
+  end.
+
+%% <stmt> ::= <uinteger> . ":"
+stmt_after_uinteger(ScanState, UInt) ->
+  case scan:token(ScanState) of
+    {ok, ?T_COLON} -> {ok, #s_local_label{number = UInt}};
     ScanRes -> badtok(ScanState, "junk after symbol", ScanRes)
   end.
 
@@ -118,6 +128,11 @@ insn_symbol(ScanState, Name, Symbol2) ->
   Displacement = #e_symbol{name = Symbol2},
   insn_ea_disp(ScanState, Name, _AccOrDev = false, _At = false, Displacement).
 
+%% Seen "<symbol> <local label>".  The <local label> is (the start of) the <displacement>.
+insn_local_label(ScanState, Name, Number, Direction) ->
+  Displacement = #e_local_label{number = Number, direction = Direction},
+  insn_ea_disp(ScanState, Name, _AccOrDev = false, _At = false, Displacement).
+
 %% <symbol> <accordev> "," . [ ["@"] <displacement> ["(" <index> ")"] ] <newline>
 insn_ea(ScanState, Name, AccOrDev) ->
   case scan:token(ScanState) of
@@ -130,6 +145,9 @@ insn_ea(ScanState, Name, AccOrDev) ->
     {ok, {?T_SYMBOL, Symbol}} ->
       Displacement = #e_symbol{name = Symbol},
       insn_ea_disp(ScanState, Name, AccOrDev, _At = false, Displacement);
+    {ok, {?T_LOCAL_LABEL, Number, Direction}} ->
+      Displacement = #e_local_label{number = Number, direction = Direction},
+      insn_ea_disp(ScanState, Name, AccOrDev, _At = false, Displacement);
     ScanRes -> badtok(ScanState, "junk after comma", ScanRes)
   end.
 
@@ -142,6 +160,9 @@ insn_ea_at(ScanState, Name, AccOrDev) ->
     {ok, {?T_SYMBOL, Symbol}} ->
       Displacement = #e_symbol{name = Symbol},
       insn_ea_disp(ScanState, Name, AccOrDev, _At = true, Displacement);
+    {ok, {?T_LOCAL_LABEL, Number, Direction}} ->
+      Displacement = #e_local_label{number = Number, direction = Direction},
+      insn_ea_disp(ScanState, Name, AccOrDev, _At = true, Displacement);
     ScanRes -> badtok(ScanState, "junk after @", ScanRes)
   end.
 
diff --git a/erlang/apps/as/src/scan.erl b/erlang/apps/as/src/scan.erl
index a38507c..a807ad5 100644
--- a/erlang/apps/as/src/scan.erl
+++ b/erlang/apps/as/src/scan.erl
@@ -206,8 +206,9 @@ do_number(ScanState, Base, Val) ->
       case chval(Ch) of
         ChVal when ChVal < Base ->
           do_number(ScanState, Base, Val * Base + ChVal);
+        _ChVal when Base =< 10 andalso (Ch =:= $b orelse Ch =:= $f) ->
+          {ok, {?T_LOCAL_LABEL, Val, Ch}};
         _ChVal ->
-          %% TODO: check for <decimal>[bf] which is a local label reference
           case scan_state:ungetc(Ch, ScanState) of
             {error, _Reason} = Error -> Error;
             ok -> {ok, {?T_UINTEGER, Val}}
diff --git a/erlang/apps/as/src/token.erl b/erlang/apps/as/src/token.erl
index ec4e9d9..9817798 100644
--- a/erlang/apps/as/src/token.erl
+++ b/erlang/apps/as/src/token.erl
@@ -48,8 +48,10 @@ format(Token) ->
     ?T_DOT_TEXT         -> ".text";
     ?T_DOT_TYPE         -> ".type";
     {?T_SYMBOL, Name}   -> io_lib:format("symbol:~s", [Name]);
-    {?T_UINTEGER, Int}  -> io_lib:format("uinteger:~p", [Int]);
+    {?T_LOCAL_LABEL, Number, Direction} ->
+      io_lib:format("label: ~.10b~c", [Number, Direction]);
     {?T_STRING, Str}    -> io_lib:format("string:~p", [Str]); % FIXME: quoting
+    {?T_UINTEGER, Int}  -> io_lib:format("uinteger:~p", [Int]);
     ?T_AT               -> "@";
     ?T_COLON            -> ":";
     ?T_COMMA            -> ",";
diff --git a/erlang/apps/as/src/token.hrl b/erlang/apps/as/src/token.hrl
index c28d6b0..00e9955 100644
--- a/erlang/apps/as/src/token.hrl
+++ b/erlang/apps/as/src/token.hrl
@@ -37,8 +37,9 @@
 -define(T_SYMBOL,       'T_SYMBOL').    % pushj, foo, .Lbar
 
 %% literals
--define(T_UINTEGER,     'T_UINTEGER').  % 017
+-define(T_LOCAL_LABEL,  'T_LOCAL_LABEL'). % 1f, 0b
 -define(T_STRING,       'T_STRING').    % "foo"
+-define(T_UINTEGER,     'T_UINTEGER').  % 017
 
 %% special symbols including operators and separators
 -define(T_AT,           'T_AT').        % @
@@ -60,8 +61,9 @@
                | ?T_DOT_TEXT
                | ?T_DOT_TYPE
                | {?T_SYMBOL, string()}
-               | {?T_UINTEGER, non_neg_integer()}
+               | {?T_LOCAL_LABEL, non_neg_integer(), $b | $f}
                | {?T_STRING, string()}
+               | {?T_UINTEGER, non_neg_integer()}
                | ?T_AT
                | ?T_COLON
                | ?T_COMMA
diff --git a/erlang/apps/as/src/tunit.erl b/erlang/apps/as/src/tunit.erl
index c44d4f6..fe266f5 100644
--- a/erlang/apps/as/src/tunit.erl
+++ b/erlang/apps/as/src/tunit.erl
@@ -25,6 +25,8 @@
         , put_section/2
         , get_symbol/2
         , put_symbol/2
+        , get_local_label/2
+        , put_local_label/3
         ]).
 
 -include("tunit.hrl").
@@ -36,6 +38,7 @@ new() ->
   #tunit{ sections = #{}
         , cursect = false
         , symbols = #{}
+        , local_labels = #{}
         }.
 
 -spec get_section(#tunit{}, string()) -> #section{} | false.
@@ -53,3 +56,11 @@ get_symbol(#tunit{symbols = Symbols}, Name) ->
 -spec put_symbol(#tunit{}, #symbol{}) -> #tunit{}.
 put_symbol(Tunit = #tunit{symbols = Symbols}, Symbol) ->
   Tunit#tunit{symbols = maps:put(Symbol#symbol.name, Symbol, Symbols)}.
+
+-spec get_local_label(#tunit{}, non_neg_integer()) -> pos_integer() | false.
+get_local_label(#tunit{local_labels = LocalLabels}, Number) ->
+  maps:get(Number, LocalLabels, false).
+
+-spec put_local_label(#tunit{}, non_neg_integer(), pos_integer()) -> #tunit{}.
+put_local_label(Tunit = #tunit{local_labels = LocalLabels}, Number, Serial) ->
+  Tunit#tunit{local_labels = maps:put(Number, Serial, LocalLabels)}.
diff --git a/erlang/apps/as/src/tunit.hrl b/erlang/apps/as/src/tunit.hrl
index 37ad674..e82e4db 100644
--- a/erlang/apps/as/src/tunit.hrl
+++ b/erlang/apps/as/src/tunit.hrl
@@ -24,9 +24,11 @@
 %% An expression occurring in a statement. (TODO: extend)
 
 -record(e_integer, {value :: integer()}).
+-record(e_local_label, {number :: non_neg_integer(), direction :: $b | $f}).
 -record(e_symbol, {name :: string()}).
 
 -type expr() :: #e_integer{}
+              | #e_local_label{}
               | #e_symbol{}
               .
 
@@ -52,8 +54,9 @@
 %% .type foo,@function (TODO: extend)
 -record(s_dot_type, {name :: string()}).
 
-%% foo:
+%% foo: 1:
 -record(s_label, {name :: string()}).
+-record(s_local_label, {number :: non_neg_integer()}).
 
 %% opcode accumulator,@address(index)
 -record(s_insn,
@@ -70,6 +73,7 @@
               | #s_dot_text{}
               | #s_dot_type{}
               | #s_label{}
+              | #s_local_label{}
               | #s_insn{}
               .
 
@@ -120,6 +124,7 @@
         { sections      :: #{string() => #section{}}
         , cursect       :: string()
         , symbols       :: #{string() => #symbol{}}
+        , local_labels  :: #{non_neg_integer() => pos_integer()}
         }).
 
 -endif. % TUNIT_HRL