[MachO] Begin to add some MachO specific File/Atoms, and add the start of

normalizedToAtoms.

llvm-svn: 198459

9 files changed, 234 insertions, 7 deletions

diff --git a/lld/lib/ReaderWriter/MachO/Atoms.h b/lld/lib/ReaderWriter/MachO/Atoms.h
new file mode 100644
index 00000000000..7733b2a9546
--- /dev/null
+++ b/lld/lib/ReaderWriter/MachO/Atoms.h
@@ -0,0 +1,41 @@
+//===- lib/ReaderWriter/MachO/Atoms.h -------------------------------------===//
+//
+//                             The LLVM Linker
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLD_READER_WRITER_MACHO_ATOMS_H
+#define LLD_READER_WRITER_MACHO_ATOMS_H
+
+#include "lld/ReaderWriter/Simple.h"
+
+namespace lld {
+namespace mach_o {
+class MachODefinedAtom : public SimpleDefinedAtom {
+public:
+  // FIXME: This constructor should also take the ContentType.
+  MachODefinedAtom(const File &f, const StringRef name,
+                   const ArrayRef<uint8_t> content)
+      : SimpleDefinedAtom(f), _name(name), _content(content) {}
+
+  virtual uint64_t size() const { return rawContent().size(); }
+
+  virtual ContentType contentType() const { return DefinedAtom::typeCode; }
+
+  virtual StringRef name() const { return _name; }
+
+  virtual Scope scope() const { return scopeGlobal; }
+
+  virtual ArrayRef<uint8_t> rawContent() const { return _content; }
+
+private:
+  const StringRef _name;
+  const ArrayRef<uint8_t> _content;
+};
+} // mach_o
+} // lld
+
+#endif
diff --git a/lld/lib/ReaderWriter/MachO/CMakeLists.txt b/lld/lib/ReaderWriter/MachO/CMakeLists.txt
index d63b4d454d7..0b9b13545e0 100644
--- a/lld/lib/ReaderWriter/MachO/CMakeLists.txt
+++ b/lld/lib/ReaderWriter/MachO/CMakeLists.txt
@@ -3,6 +3,7 @@ add_lld_library(lldMachO
   MachONormalizedFileBinaryReader.cpp
   MachONormalizedFileBinaryWriter.cpp
   MachONormalizedFileFromAtoms.cpp
+  MachONormalizedFileToAtoms.cpp
   MachONormalizedFileYAML.cpp
   ReferenceKinds.cpp
   WriterMachO.cpp
@@ -12,3 +13,5 @@ target_link_libraries(lldMachO
   lldReaderWriter
   lldPasses
   )
+
+include_directories(.)
diff --git a/lld/lib/ReaderWriter/MachO/File.h b/lld/lib/ReaderWriter/MachO/File.h
new file mode 100644
index 00000000000..e002a93e4b4
--- /dev/null
+++ b/lld/lib/ReaderWriter/MachO/File.h
@@ -0,0 +1,37 @@
+//===- lib/ReaderWriter/MachO/File.h --------------------------------------===//
+//
+//                             The LLVM Linker
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLD_READER_WRITER_MACHO_FILE_H
+#define LLD_READER_WRITER_MACHO_FILE_H
+
+#include "Atoms.h"
+
+#include "lld/ReaderWriter/Simple.h"
+
+namespace lld {
+namespace mach_o {
+
+class MachOFile : public SimpleFile {
+public:
+  MachOFile(StringRef path) : SimpleFile(path) {}
+
+  void addDefinedAtom(StringRef name, ArrayRef<uint8_t> content) {
+    MachODefinedAtom *atom =
+        new (_allocator) MachODefinedAtom(*this, name, content);
+    addAtom(*atom);
+  }
+
+private:
+  llvm::BumpPtrAllocator _allocator;
+};
+
+} // end namespace mach_o
+} // end namespace lld
+
+#endif
diff --git a/lld/lib/ReaderWriter/MachO/MachONormalizedFile.h b/lld/lib/ReaderWriter/MachO/MachONormalizedFile.h
index f2f6cf86184..15fd4c3686d 100644
--- a/lld/lib/ReaderWriter/MachO/MachONormalizedFile.h
+++ b/lld/lib/ReaderWriter/MachO/MachONormalizedFile.h
@@ -113,7 +113,7 @@ struct Section {
   SectionAttr     attributes;
   uint32_t        alignment;
   Hex64           address;
-  ContentBytes    content;
+  std::vector<uint8_t> content;
   Relocations     relocations;
   IndirectSymbols indirectSymbols;
 };
@@ -273,7 +273,3 @@ normalizedFromAtoms(const lld::File &atomFile, const MachOLinkingContext &ctxt);
 } // namespace lld
 
 #endif // LLD_READER_WRITER_MACHO_NORMALIZE_FILE_H
-
-
-
-
diff --git a/lld/lib/ReaderWriter/MachO/MachONormalizedFileFromAtoms.cpp b/lld/lib/ReaderWriter/MachO/MachONormalizedFileFromAtoms.cpp
index 0d73ec81180..1479ac20da2 100644
--- a/lld/lib/ReaderWriter/MachO/MachONormalizedFileFromAtoms.cpp
+++ b/lld/lib/ReaderWriter/MachO/MachONormalizedFileFromAtoms.cpp
@@ -423,7 +423,7 @@ void Util::appendSection(SectionInfo *si, NormalizedFile &file) {
   // Copy content from atoms to content buffer for section.
   // FIXME: zerofill atoms/sections should not take up content space.
   normSect->content.resize(si->size);
-  Hex8 *sectionContent = normSect->content.data();
+  uint8_t *sectionContent = normSect->content.data();
   for (AtomInfo &ai : si->atomsAndOffsets) {
     // Copy raw bytes.
     uint8_t *atomContent = reinterpret_cast<uint8_t*>
diff --git a/lld/lib/ReaderWriter/MachO/MachONormalizedFileToAtoms.cpp b/lld/lib/ReaderWriter/MachO/MachONormalizedFileToAtoms.cpp
new file mode 100644
index 00000000000..2032e0cea05
--- /dev/null
+++ b/lld/lib/ReaderWriter/MachO/MachONormalizedFileToAtoms.cpp
@@ -0,0 +1,66 @@
+//===- lib/ReaderWriter/MachO/MachONormalizedFileToAtoms.cpp --------------===//
+//
+//                             The LLVM Linker
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+///
+/// \file Converts from in-memory normalized mach-o to in-memory Atoms.
+///
+///                  +------------+
+///                  | normalized |
+///                  +------------+
+///                        |
+///                        |
+///                        v
+///                    +-------+
+///                    | Atoms |
+///                    +-------+
+
+#include "MachONormalizedFile.h"
+#include "File.h"
+#include "Atoms.h"
+
+#include "lld/Core/LLVM.h"
+
+#include "llvm/Support/MachO.h"
+
+using namespace llvm::MachO;
+
+namespace lld {
+namespace mach_o {
+namespace normalized {
+
+static ErrorOr<std::unique_ptr<lld::File>>
+normalizedObjectToAtoms(const NormalizedFile &normalizedFile, StringRef path) {
+  std::unique_ptr<MachOFile> file(new MachOFile(path));
+
+  for (auto &sym : normalizedFile.globalSymbols) {
+    file->addDefinedAtom(sym.name,
+                         normalizedFile.sections[sym.sect - 1].content);
+  }
+
+  assert(normalizedFile.localSymbols.empty() &&
+         "local symbols not supported yet!");
+  assert(normalizedFile.undefinedSymbols.empty() &&
+         "undefined symbols not supported yet!");
+
+  return std::unique_ptr<File>(std::move(file));
+}
+
+ErrorOr<std::unique_ptr<lld::File>>
+normalizedToAtoms(const NormalizedFile &normalizedFile, StringRef path) {
+  switch (normalizedFile.fileType) {
+  case MH_OBJECT:
+    return normalizedObjectToAtoms(normalizedFile, path);
+  default:
+    llvm_unreachable("unhandled MachO file type!");
+  }
+}
+
+} // namespace normalized
+} // namespace mach_o
+} // namespace lld
diff --git a/lld/lib/ReaderWriter/MachO/MachONormalizedFileYAML.cpp b/lld/lib/ReaderWriter/MachO/MachONormalizedFileYAML.cpp
index 2a888d5d3eb..78efae6adb8 100644
--- a/lld/lib/ReaderWriter/MachO/MachONormalizedFileYAML.cpp
+++ b/lld/lib/ReaderWriter/MachO/MachONormalizedFileYAML.cpp
@@ -269,6 +269,7 @@ struct ScalarBitSetTraits<SymbolDesc> {
 
 template <>
 struct MappingTraits<Section> {
+  struct NormalizedContentBytes;
   static void mapping(IO &io, Section &sect) {
     io.mapRequired("segment",         sect.segmentName);
     io.mapRequired("section",         sect.sectionName);
@@ -276,10 +277,31 @@ struct MappingTraits<Section> {
     io.mapOptional("attributes",      sect.attributes);
     io.mapOptional("alignment",       sect.alignment, 0U);
     io.mapRequired("address",         sect.address);
-    io.mapOptional("content",         sect.content);
+    MappingNormalization<NormalizedContent, std::vector<uint8_t>> content(
+        io, sect.content);
+    io.mapOptional("content",         content->normalizedContent);
     io.mapOptional("relocations",     sect.relocations);
     io.mapOptional("indirect-syms",   sect.indirectSymbols);
   }
+
+  // FIXME: It would be good if we could remove this, so we don't need to copy
+  // the content data.
+  struct NormalizedContent {
+    NormalizedContent(IO &) {}
+    NormalizedContent(IO &, std::vector<uint8_t> content) {
+      for (auto &c : content) {
+        normalizedContent.push_back(c);
+      }
+    }
+    std::vector<uint8_t> denormalize(IO &) {
+      std::vector<uint8_t> content;
+      for (auto &c : normalizedContent) {
+        content.push_back(c);
+      }
+      return content;
+    }
+    ContentBytes normalizedContent;
+  };
 };
 
 
diff --git a/lld/unittests/MachOTests/CMakeLists.txt b/lld/unittests/MachOTests/CMakeLists.txt
index e95d261ff7c..23052d366c6 100644
--- a/lld/unittests/MachOTests/CMakeLists.txt
+++ b/lld/unittests/MachOTests/CMakeLists.txt
@@ -2,6 +2,7 @@
 add_lld_unittest(lldMachOTests
   MachONormalizedFileBinaryReaderTests.cpp
   MachONormalizedFileBinaryWriterTests.cpp
+  MachONormalizedFileToAtomsTests.cpp
   MachONormalizedFileYAMLTests.cpp
   )
 
diff --git a/lld/unittests/MachOTests/MachONormalizedFileToAtomsTests.cpp b/lld/unittests/MachOTests/MachONormalizedFileToAtomsTests.cpp
new file mode 100644
index 00000000000..9b2859a5961
--- /dev/null
+++ b/lld/unittests/MachOTests/MachONormalizedFileToAtomsTests.cpp
@@ -0,0 +1,61 @@
+//===- lld/unittest/MachOTests/MachONormalizedFileToAtomsTests.cpp --------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#include "gtest/gtest.h"
+
+#include <llvm/Support/MachO.h>
+#include "../../lib/ReaderWriter/MachO/MachONormalizedFile.h"
+
+#include <assert.h>
+#include <vector>
+
+using llvm::ErrorOr;
+
+using namespace lld::mach_o::normalized;
+using namespace llvm::MachO;
+
+unsigned countDefinedAtoms(const lld::File &file) {
+  unsigned count = 0;
+  for (const auto &d : file.defined()) {
+    (void)d;
+    ++count;
+  }
+  return count;
+}
+
+TEST(ToAtomsTest, empty_obj_x86_64) {
+  NormalizedFile f;
+  f.arch = lld::MachOLinkingContext::arch_x86_64;
+  ErrorOr<std::unique_ptr<const lld::File>> atom_f = normalizedToAtoms(f, "");
+  EXPECT_FALSE(!atom_f);
+  EXPECT_EQ(0U, countDefinedAtoms(**atom_f));
+}
+
+TEST(ToAtomsTest, basic_obj_x86_64) {
+  NormalizedFile f;
+  f.arch = lld::MachOLinkingContext::arch_x86_64;
+  Section textSection;
+  static const uint8_t contentBytes[] = { 0x55, 0x48, 0x89, 0xE5,
+                                          0x31, 0xC0, 0x5D, 0xC3 };
+  const unsigned contentSize = sizeof(contentBytes) / sizeof(contentBytes[0]);
+  textSection.content.insert(textSection.content.begin(), contentBytes,
+                             &contentBytes[contentSize]);
+  f.sections.push_back(textSection);
+  Symbol mainSymbol;
+  mainSymbol.name = "_main";
+  mainSymbol.type = N_SECT;
+  mainSymbol.sect = 1;
+  f.globalSymbols.push_back(mainSymbol);
+  ErrorOr<std::unique_ptr<const lld::File>> atom_f = normalizedToAtoms(f, "");
+  EXPECT_FALSE(!atom_f);
+  EXPECT_EQ(1U, countDefinedAtoms(**atom_f));
+  const lld::DefinedAtom *singleAtom = *(*atom_f)->defined().begin();
+  llvm::ArrayRef<uint8_t> atomContent(singleAtom->rawContent());
+  EXPECT_EQ(0, memcmp(atomContent.data(), contentBytes, contentSize));
+}