From f6096ec45c2da3bf87f5669db1175b3d9721a7b2 Mon Sep 17 00:00:00 2001
From: Mikael Pettersson <mikpelinux@gmail.com>
Date: Wed, 9 Aug 2023 20:49:05 +0200
Subject: [PATCH] ld_input: add support for archives

---
 erlang/apps/ld/src/ld_input.erl    | 67 ++++++++++++++++++++++++++++--
 erlang/apps/ld/src/ld_internal.hrl | 11 ++++-
 2 files changed, 72 insertions(+), 6 deletions(-)

diff --git a/erlang/apps/ld/src/ld_input.erl b/erlang/apps/ld/src/ld_input.erl
index 9549298..7bab34c 100644
--- a/erlang/apps/ld/src/ld_input.erl
+++ b/erlang/apps/ld/src/ld_input.erl
@@ -17,6 +17,19 @@
 %%%
 %%% You should have received a copy of the GNU General Public License
 %%% along with pdp10-tools.  If not, see <http://www.gnu.org/licenses/>.
+%%%
+%%%=============================================================================
+%%%
+%%% - Process input files in the order given.
+%%% - Maintain sets of currently defined and undefined symbols, initially empty
+%%%   and the entry symbol, respectiively.
+%%% - An input object file is included unconditionally. Its symbol table is read
+%%%   and the sets of defined and undefined symbols are updated accordingly.
+%%% - An input archive has its members processed in the order stored.
+%%% - An input archive member is included if any only if the archive's symbol table
+%%%   indicates that the member defines some currently undefined symbol.
+%%% - If an archive member is included it is processed exactly like an explicit
+%%%   input object file.
 
 -module(ld_input).
 
@@ -25,6 +38,8 @@
         ]).
 
 -include("ld_internal.hrl").
+-include_lib("lib/include/archive.hrl").
+-include_lib("lib/include/pdp10_ar.hrl").
 
 %% error reasons
 -define(badelf, badelf).
@@ -55,14 +70,58 @@ input([], _DefMap, UndefMap, Inputs) ->
   end.
 
 input_file(File, DefMap, UndefMap, Inputs) ->
-  case pdp10_stdio:fopen(File, [raw, read]) of
-    {ok, FP} ->
+  case archive:read(File) of
+    {ok, {FP, Archive}} ->
       try
-        input_elf(File, FP, DefMap, UndefMap, Inputs)
+        input_archive(File, FP, Archive, DefMap, UndefMap, Inputs)
       after
         pdp10_stdio:fclose(FP)
       end;
-    {error, Reason} -> mkerror({?badfile, File, Reason})
+    {error, _} ->
+      case pdp10_stdio:fopen(File, [raw, read]) of
+        {ok, FP} ->
+          try
+            input_elf(File, FP, DefMap, UndefMap, Inputs)
+          after
+            pdp10_stdio:fclose(FP)
+          end;
+        {error, Reason} -> mkerror({?badfile, File, Reason})
+      end
+  end.
+
+%% Process input archive =======================================================
+
+input_archive(File, FP, Archive, DefMap, UndefMap, Inputs) ->
+  #archive{symtab = SymTab, members = Members} = Archive,
+  %% Build a mapping from member offset to the set of symbols defined there.
+  OffsetToDefs =
+    lists:foldl(
+      fun({Symbol, Offset}, Map) ->
+        Symbols = maps:get(Offset, Map, #{}),
+        maps:put(Offset, maps:put(Symbol, false, Symbols), Map)
+      end, maps:new(), SymTab),
+  input_archive(Members, OffsetToDefs, File, FP, DefMap, UndefMap, Inputs).
+
+input_archive(_Members = [], _OffsetToDefs, _File, _FP, DefMap, UndefMap, Inputs) ->
+  {ok, {DefMap, UndefMap, Inputs}};
+input_archive([Member | Members], OffsetToDefs, File, FP, DefMap, UndefMap, Inputs) ->
+  case input_member(Member, OffsetToDefs, File, FP, DefMap, UndefMap) of
+    false ->
+      input_archive(Members, OffsetToDefs, File, FP, DefMap, UndefMap, Inputs);
+    {ok, {NewDefMap, NewUndefMap, Input}} ->
+      input_archive(Members, OffsetToDefs, File, FP, NewDefMap, NewUndefMap, [Input | Inputs]);
+    {error, _Reason} = Error -> Error
+  end.
+
+input_member(Member, OffsetToDefs, File, FP, DefMap, UndefMap) ->
+  #member{arhdr = ArHdr, location = HdrOffset} = Member,
+  MemberDefs = maps:get(HdrOffset, OffsetToDefs, #{}),
+  case maps:size(maps:intersect(MemberDefs, UndefMap)) of
+    0 -> false;
+    _ ->
+      Offset = HdrOffset + ?PDP10_ARHDR_SIZEOF,
+      #arhdr{ar_name = Name, ar_size = Size} = ArHdr,
+      input_elf({File, Name, Offset, Size}, FP, Offset, Offset + Size, DefMap, UndefMap)
   end.
 
 %% Process input object file ===================================================
diff --git a/erlang/apps/ld/src/ld_internal.hrl b/erlang/apps/ld/src/ld_internal.hrl
index 8206f40..ba954d9 100644
--- a/erlang/apps/ld/src/ld_internal.hrl
+++ b/erlang/apps/ld/src/ld_internal.hrl
@@ -1,7 +1,7 @@
 %%% -*- erlang-indent-level: 2 -*-
 %%%
 %%% internal declarations for pdp10-elf ld.
-%%% Copyright (C) 2020  Mikael Pettersson
+%%% Copyright (C) 2020-2023  Mikael Pettersson
 %%%
 %%% This file is part of pdp10-tools.
 %%%
@@ -23,8 +23,15 @@
 
 -include_lib("lib/include/pdp10_elf36.hrl").
 
+-type ifile() :: string() % explicit input .o-file
+               | { Archive :: string() % archive name
+                 , Name :: string() % member name
+                 , Offset :: non_neg_integer() % member data (not header) at this offset
+                 , Size :: non_neg_integer() % member data size
+                 }.
+
 -record(input,
-        { file    :: string()
+        { file    :: ifile()
         , shtab   :: [#elf36_Shdr{}]
         , symtab  :: [#elf36_Sym{}]
         , stshndx :: non_neg_integer()