From 8f0acbf413463592adfcb7f4d7f58848fb8c7324 Mon Sep 17 00:00:00 2001
From: Rui Ueyama <ruiu@cs.stanford.edu>
Date: Thu, 17 Feb 2022 14:22:06 +0900
Subject: [PATCH] [ELF][LTO] Move LTO to the symbol resolution pass

ELF object files returned by do_lto() may contain unforeseen undefined
symbols, so we can't call that function after the symbol resolution pass
because it can cause undefined symbol errors.
---
 elf/lto.cc    |  7 ------
 elf/main.cc   |  4 ----
 elf/passes.cc | 66 ++++++++++++++++++++++++++++++++++-----------------
 3 files changed, 44 insertions(+), 33 deletions(-)

diff --git a/elf/lto.cc b/elf/lto.cc
index 11048cd949..c35c379052 100644
--- a/elf/lto.cc
+++ b/elf/lto.cc
@@ -163,7 +163,6 @@ static PluginStatus add_input_file(const char *path) {
   file->is_alive = true;
   file->parse(ctx);
   file->resolve_symbols(ctx);
-  file->mark_live_objects(ctx, [](InputFile<E> *) { unreachable(); });
   return LDPS_OK;
 }
 
@@ -538,12 +537,6 @@ void do_lto(Context<E> &ctx) {
   assert(phase == 1);
   phase = 2;
 
-  // Compute import/export information early because `get_symbols`
-  // function needs them.
-  apply_version_script(ctx);
-  parse_symbol_version(ctx);
-  compute_import_export(ctx);
-
   // Set `referenced_by_regular_obj` bit.
   for (ObjectFile<E> *file : ctx.objs) {
     if (file->is_lto_obj)
diff --git a/elf/main.cc b/elf/main.cc
index 9b80cd49f0..103206ad91 100644
--- a/elf/main.cc
+++ b/elf/main.cc
@@ -469,10 +469,6 @@ static int elf_main(int argc, char **argv) {
   // included to the final output.
   resolve_symbols(ctx);
 
-  // Do LTO
-  if (ctx.has_lto_object)
-    do_lto(ctx);
-
   // Resolve mergeable section pieces to merge them.
   register_section_pieces(ctx);
 
diff --git a/elf/passes.cc b/elf/passes.cc
index 03ed7496a8..3a14808d0b 100644
--- a/elf/passes.cc
+++ b/elf/passes.cc
@@ -77,27 +77,10 @@ void create_synthetic_sections(Context<E> &ctx) {
 }
 
 template <typename E>
-void resolve_symbols(Context<E> &ctx) {
-  Timer t(ctx, "resolve_symbols");
-
-  std::vector<InputFile<E> *> files;
-  append(files, ctx.objs);
-  append(files, ctx.dsos);
-
-  // Register symbols
-  tbb::parallel_for_each(files, [&](InputFile<E> *file) {
-    file->resolve_symbols(ctx);
-  });
-
-  // Mark reachable objects to decide which files to include
-  // into an output.
-  std::vector<InputFile<E> *> live_set = files;
-  std::erase_if(live_set, [](InputFile<E> *file) { return !file->is_alive; });
-
+static void mark_live_objects(Context<E> &ctx) {
   auto mark_symbol = [&](std::string_view name) {
     if (InputFile<E> *file = get_symbol(ctx, name)->file)
-      if (!file->is_alive.exchange(true))
-        live_set.push_back(file);
+      file->is_alive = true;
   };
 
   for (std::string_view name : ctx.arg.undefined)
@@ -105,11 +88,50 @@ void resolve_symbols(Context<E> &ctx) {
   for (std::string_view name : ctx.arg.require_defined)
     mark_symbol(name);
 
-  tbb::parallel_for_each(live_set, [&](InputFile<E> *file,
-                                       tbb::feeder<InputFile<E> *> &feeder) {
-    file->mark_live_objects(ctx, [&](InputFile<E> *obj) { feeder.add(obj); });
+  auto mark_file = [&](InputFile<E> *file, tbb::feeder<InputFile<E> *> &feeder) {
+    if (file->is_alive)
+      file->mark_live_objects(ctx, [&](InputFile<E> *obj) { feeder.add(obj); });
+  };
+
+  tbb::parallel_for_each(ctx.objs, mark_file);
+  tbb::parallel_for_each(ctx.dsos, mark_file);
+}
+
+template <typename E>
+void resolve_symbols(Context<E> &ctx) {
+  Timer t(ctx, "resolve_symbols");
+
+  // Register symbols
+  tbb::parallel_for_each(ctx.objs, [&](InputFile<E> *file) {
+    file->resolve_symbols(ctx);
   });
 
+  tbb::parallel_for_each(ctx.dsos, [&](InputFile<E> *file) {
+    file->resolve_symbols(ctx);
+  });
+
+  // Do link-time optimization. We pass all IR object files to the
+  // compiler backend to compile them into a few ELF object files.
+  if (ctx.has_lto_object) {
+    Timer t(ctx, "do_lto");
+
+    // The compiler backend needs to know how symbols are resolved,
+    // so compute symbolvisibility, import/export bits, etc early.
+    mark_live_objects(ctx);
+    apply_version_script(ctx);
+    parse_symbol_version(ctx);
+    compute_import_export(ctx);
+    do_lto(ctx);
+  }
+
+  // Mark reachable objects to decide which files to include into an output.
+  // This also merges symbol visibility.
+  mark_live_objects(ctx);
+
+  std::vector<InputFile<E> *> files;
+  append(files, ctx.objs);
+  append(files, ctx.dsos);
+
   // Remove symbols of eliminated files.
   tbb::parallel_for_each(files, [&](InputFile<E> *file) {
     if (!file->is_alive)