Fix memory leak caused by treating Code::next_code_link as strong in marker.

src/objects-visiting-inl.h

 // Copyright 2012 the V8 project authors. All rights reserved.
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions are
 // met:
 //
 //     * Redistributions of source code must retain the above copyright
 //       notice, this list of conditions and the following disclaimer.
 //     * Redistributions in binary form must reproduce the above
 //       copyright notice, this list of conditions and the following
 //       disclaimer in the documentation and/or other materials provided
@@ skipping 879 matching lines @@
                   RelocInfo::ModeMask(RelocInfo::EXTERNAL_REFERENCE) |
                   RelocInfo::ModeMask(RelocInfo::JS_RETURN) |
                   RelocInfo::ModeMask(RelocInfo::DEBUG_BREAK_SLOT) |
                   RelocInfo::ModeMask(RelocInfo::RUNTIME_ENTRY);
 
   // There are two places where we iterate code bodies: here and the
   // templated CodeIterateBody (below). They should be kept in sync.
   IteratePointer(v, kRelocationInfoOffset);
   IteratePointer(v, kHandlerTableOffset);
   IteratePointer(v, kDeoptimizationDataOffset);
-  IteratePointer(v, kTypeFeedbackInfoOffset);
+  if (v->weak_pointer_mode() == VISIT_WEAK_POINTERS || !is_optimized_code()) {
+    // In optimized code this field is a weak next_code_link.
+    IteratePointer(v, kTypeFeedbackInfoOffset);
+  }
   IteratePointer(v, kConstantPoolOffset);
 
   RelocIterator it(this, mode_mask);
   Isolate* isolate = this->GetIsolate();
   for (; !it.done(); it.next()) {
     it.rinfo()->Visit(isolate, v);
   }
 }
 
 
@@ skipping 11 matching lines @@
   // templated CodeIterateBody (above). They should be kept in sync.
   StaticVisitor::VisitPointer(
       heap,
       reinterpret_cast<Object**>(this->address() + kRelocationInfoOffset));
   StaticVisitor::VisitPointer(
       heap,
       reinterpret_cast<Object**>(this->address() + kHandlerTableOffset));
   StaticVisitor::VisitPointer(
       heap,
       reinterpret_cast<Object**>(this->address() + kDeoptimizationDataOffset));
-  StaticVisitor::VisitPointer(
-      heap,
-      reinterpret_cast<Object**>(this->address() + kTypeFeedbackInfoOffset));
+  if (StaticVisitor::weak_pointer_mode() == VISIT_WEAK_POINTERS ||
+      !is_optimized_code()) {

[Michael Starzinger, 2014/03/04 12:31:12]

As discussed offline: AFAICT, currently we can afford to not have the fields
overlap as the padding is big enough on 32bit and 64bit architectures. So
"un-overlapping" them woukd definitely make things easier to read.

[titzer, 2014/03/05 08:58:02]

+1 unoverlap. I am sorry for overlapping them in the first place.

[ulan, 2014/03/06 11:13:32]

Nope, it is my fault. I remember introducing this field and overlapping it.
Sorry for that.

[ulan, 2014/03/06 11:13:32]

Done.

+    // In optimized code this field is a weak next_code_link.
+    StaticVisitor::VisitPointer(
+        heap,
+        reinterpret_cast<Object**>(this->address() + kTypeFeedbackInfoOffset));
+  }
   StaticVisitor::VisitPointer(
       heap,
       reinterpret_cast<Object**>(this->address() + kConstantPoolOffset));
 
 
   RelocIterator it(this, mode_mask);
   for (; !it.done(); it.next()) {
     it.rinfo()->template Visit<StaticVisitor>(heap);
   }
 }
 
 
 } }  // namespace v8::internal
 
 #endif  // V8_OBJECTS_VISITING_INL_H_