Re-land "Fix Bug 30978 by emitting cv file checksums."

This reverts r313431 and brings back r313374 with a fix to write
checksums as binary data and not ASCII hex strings.

llvm-svn: 313657

Author: rnk

llvm/include/llvm/IR/DebugInfoMetadata.h

@@ -473,10 +473,12 @@ class DIFile : public DIScope {
   friend class MDNode;
 
 public:
+  // These values must be explictly set, as they end up in the final object
+  // file.
   enum ChecksumKind {
-    CSK_None,
-    CSK_MD5,
-    CSK_SHA1,
+    CSK_None = 0,
+    CSK_MD5 = 1,
+    CSK_SHA1 = 2,
     CSK_Last = CSK_SHA1 // Should be last enumeration.
   };
 
@@ -510,7 +512,7 @@ class DIFile : public DIScope {
                              ChecksumKind CSK = CSK_None,
                              StringRef CS = StringRef()),
                     (Filename, Directory, CSK, CS))
-  DEFINE_MDNODE_GET(DIFile, (MDString *Filename, MDString *Directory,
+  DEFINE_MDNODE_GET(DIFile, (MDString * Filename, MDString *Directory,
                              ChecksumKind CSK = CSK_None,
                              MDString *CS = nullptr),
                     (Filename, Directory, CSK, CS))

llvm/include/llvm/MC/MCCodeView.h

@@ -161,8 +161,8 @@ class CodeViewContext {
   ~CodeViewContext();
 
   bool isValidFileNumber(unsigned FileNumber) const;
-  bool addFile(unsigned FileNumber, StringRef Filename);
-  ArrayRef<StringRef> getFilenames() { return Filenames; }
+  bool addFile(MCStreamer &OS, unsigned FileNumber, StringRef Filename,
+               ArrayRef<uint8_t> ChecksumBytes, uint8_t ChecksumKind);
 
   /// Records the function id of a normal function. Returns false if the
   /// function id has already been used, and true otherwise.
@@ -273,6 +273,9 @@ class CodeViewContext {
   /// Emits the file checksum substream.
   void emitFileChecksums(MCObjectStreamer &OS);
 
+  /// Emits the offset into the checksum table of the given file number.
+  void emitFileChecksumOffset(MCObjectStreamer &OS, unsigned FileNo);
+
 private:
   /// The current CodeView line information from the last .cv_loc directive.
   MCCVLoc CurrentCVLoc = MCCVLoc(0, 0, 0, 0, false, true);
@@ -287,14 +290,31 @@ class CodeViewContext {
 
   MCDataFragment *getStringTableFragment();
 
-  /// Add something to the string table.
-  StringRef addToStringTable(StringRef S);
+  /// Add something to the string table.  Returns the final string as well as
+  /// offset into the string table.
+  std::pair<StringRef, unsigned> addToStringTable(StringRef S);
 
   /// Get a string table offset.
   unsigned getStringTableOffset(StringRef S);
 
-  /// An array of absolute paths. Eventually this may include the file checksum.
-  SmallVector<StringRef, 4> Filenames;
+  struct FileInfo {
+    unsigned StringTableOffset;
+
+    // Indicates if this FileInfo corresponds to an actual file, or hasn't been
+    // set yet.
+    bool Assigned = false;
+
+    uint8_t ChecksumKind;
+
+    ArrayRef<uint8_t> Checksum;
+
+    // Checksum offset stored as a symbol because it might be requested
+    // before it has been calculated, so a fixup may be needed.
+    MCSymbol *ChecksumTableOffset;
+  };
+
+  /// Array storing added file information.
+  SmallVector<FileInfo, 4> Files;
 
   /// The offset of the first and last .cv_loc directive for a given function
   /// id.
@@ -305,6 +325,10 @@ class CodeViewContext {
 
   /// All known functions and inlined call sites, indexed by function id.
   std::vector<MCCVFunctionInfo> Functions;
+
+  /// Indicate whether we have already laid out the checksum table addresses or
+  /// not.
+  bool ChecksumOffsetsAssigned = false;
 };
 
 } // end namespace llvm

llvm/include/llvm/MC/MCObjectStreamer.h

@@ -140,6 +140,7 @@ class MCObjectStreamer : public MCStreamer {
       StringRef FixedSizePortion) override;
   void EmitCVStringTableDirective() override;
   void EmitCVFileChecksumsDirective() override;
+  void EmitCVFileChecksumOffsetDirective(unsigned FileNo) override;
   void EmitDTPRel32Value(const MCExpr *Value) override;
   void EmitDTPRel64Value(const MCExpr *Value) override;
   void EmitTPRel32Value(const MCExpr *Value) override;

llvm/include/llvm/MC/MCStreamer.h

@@ -733,10 +733,12 @@ class MCStreamer {
                                      unsigned Isa, unsigned Discriminator,
                                      StringRef FileName);
 
-  /// \brief Associate a filename with a specified logical file number.  This
-  /// implements the '.cv_file 4 "foo.c"' assembler directive. Returns true on
-  /// success.
-  virtual bool EmitCVFileDirective(unsigned FileNo, StringRef Filename);
+  /// Associate a filename with a specified logical file number, and also
+  /// specify that file's checksum information.  This implements the '.cv_file 4
+  /// "foo.c"' assembler directive. Returns true on success.
+  virtual bool EmitCVFileDirective(unsigned FileNo, StringRef Filename,
+                                   ArrayRef<uint8_t> Checksum,
+                                   unsigned ChecksumKind);
 
   /// \brief Introduces a function id for use with .cv_loc.
   virtual bool EmitCVFuncIdDirective(unsigned FunctionId);
@@ -778,6 +780,10 @@ class MCStreamer {
   /// \brief This implements the CodeView '.cv_filechecksums' assembler directive.
   virtual void EmitCVFileChecksumsDirective() {}
 
+  /// This implements the CodeView '.cv_filechecksumoffset' assembler
+  /// directive.
+  virtual void EmitCVFileChecksumOffsetDirective(unsigned FileNo) {}
+
   /// Emit the absolute difference between two symbols.
   ///
   /// \pre Offset of \c Hi is greater than the offset \c Lo.

llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.cpp

@@ -159,7 +159,14 @@ unsigned CodeViewDebug::maybeRecordFile(const DIFile *F) {
   if (Insertion.second) {
     // We have to compute the full filepath and emit a .cv_file directive.
     StringRef FullPath = getFullFilepath(F);
-    bool Success = OS.EmitCVFileDirective(NextId, FullPath);
+    std::string Checksum = fromHex(F->getChecksum());
+    void *CKMem = OS.getContext().allocate(Checksum.size(), 1);
+    memcpy(CKMem, Checksum.data(), Checksum.size());
+    ArrayRef<uint8_t> ChecksumAsBytes(reinterpret_cast<const uint8_t *>(CKMem),
+                                      Checksum.size());
+    DIFile::ChecksumKind ChecksumKind = F->getChecksumKind();
+    bool Success = OS.EmitCVFileDirective(NextId, FullPath, ChecksumAsBytes,
+                                          static_cast<unsigned>(ChecksumKind));
     (void)Success;
     assert(Success && ".cv_file directive failed");
   }
@@ -681,8 +688,10 @@ void CodeViewDebug::emitInlineeLinesSubsection() {
   OS.AddComment("Inlinee lines subsection");
   MCSymbol *InlineEnd = beginCVSubsection(DebugSubsectionKind::InlineeLines);
 
-  // We don't provide any extra file info.
-  // FIXME: Find out if debuggers use this info.
+  // We emit the checksum info for files.  This is used by debuggers to
+  // determine if a pdb matches the source before loading it.  Visual Studio,
+  // for instance, will display a warning that the breakpoints are not valid if
+  // the pdb does not match the source.
   OS.AddComment("Inlinee lines signature");
   OS.EmitIntValue(unsigned(InlineeLinesSignature::Normal), 4);
 
@@ -695,13 +704,10 @@ void CodeViewDebug::emitInlineeLinesSubsection() {
     OS.AddComment("Inlined function " + SP->getName() + " starts at " +
                   SP->getFilename() + Twine(':') + Twine(SP->getLine()));
     OS.AddBlankLine();
-    // The filechecksum table uses 8 byte entries for now, and file ids start at
-    // 1.
-    unsigned FileOffset = (FileId - 1) * 8;
     OS.AddComment("Type index of inlined function");
     OS.EmitIntValue(InlineeIdx.getIndex(), 4);
     OS.AddComment("Offset into filechecksum table");
-    OS.EmitIntValue(FileOffset, 4);
+    OS.EmitCVFileChecksumOffsetDirective(FileId);
     OS.AddComment("Starting line number");
     OS.EmitIntValue(SP->getLine(), 4);
   }

llvm/lib/IR/DebugInfoMetadata.cpp

@@ -354,6 +354,8 @@ DISubroutineType *DISubroutineType::getImpl(LLVMContext &Context, DIFlags Flags,
   DEFINE_GETIMPL_STORE(DISubroutineType, (Flags, CC), Ops);
 }
 
+// FIXME: Implement this string-enum correspondence with a .def file and macros,
+// so that the association is explicit rather than implied.
 static const char *ChecksumKindName[DIFile::CSK_Last + 1] = {
   "CSK_None",
   "CSK_MD5",

llvm/lib/MC/MCAsmStreamer.cpp

@@ -225,7 +225,9 @@ class MCAsmStreamer final : public MCStreamer {
                              StringRef FileName) override;
   MCSymbol *getDwarfLineTableSymbol(unsigned CUID) override;
 
-  bool EmitCVFileDirective(unsigned FileNo, StringRef Filename) override;
+  bool EmitCVFileDirective(unsigned FileNo, StringRef Filename,
+                           ArrayRef<uint8_t> Checksum,
+                           unsigned ChecksumKind) override;
   bool EmitCVFuncIdDirective(unsigned FuncId) override;
   bool EmitCVInlineSiteIdDirective(unsigned FunctionId, unsigned IAFunc,
                                    unsigned IAFile, unsigned IALine,
@@ -245,6 +247,7 @@ class MCAsmStreamer final : public MCStreamer {
       StringRef FixedSizePortion) override;
   void EmitCVStringTableDirective() override;
   void EmitCVFileChecksumsDirective() override;
+  void EmitCVFileChecksumOffsetDirective(unsigned FileNo) override;
 
   void EmitIdent(StringRef IdentString) override;
   void EmitCFISections(bool EH, bool Debug) override;
@@ -1120,13 +1123,25 @@ MCSymbol *MCAsmStreamer::getDwarfLineTableSymbol(unsigned CUID) {
   return MCStreamer::getDwarfLineTableSymbol(0);
 }
 
-bool MCAsmStreamer::EmitCVFileDirective(unsigned FileNo, StringRef Filename) {
-  if (!getContext().getCVContext().addFile(FileNo, Filename))
+bool MCAsmStreamer::EmitCVFileDirective(unsigned FileNo, StringRef Filename,
+                                        ArrayRef<uint8_t> Checksum,
+                                        unsigned ChecksumKind) {
+  if (!getContext().getCVContext().addFile(*this, FileNo, Filename, Checksum,
+                                           ChecksumKind))
     return false;
 
   OS << "\t.cv_file\t" << FileNo << ' ';
-
   PrintQuotedString(Filename, OS);
+
+  if (!ChecksumKind) {
+    EmitEOL();
+    return true;
+  }
+
+  OS << ' ';
+  PrintQuotedString(toHex(Checksum), OS);
+  OS << ' ' << ChecksumKind;
+
   EmitEOL();
   return true;
 }
@@ -1228,6 +1243,11 @@ void MCAsmStreamer::EmitCVFileChecksumsDirective() {
   EmitEOL();
 }
 
+void MCAsmStreamer::EmitCVFileChecksumOffsetDirective(unsigned FileNo) {
+  OS << "\t.cv_filechecksumoffset\t" << FileNo;
+  EmitEOL();
+}
+
 void MCAsmStreamer::EmitIdent(StringRef IdentString) {
   assert(MAI->hasIdentDirective() && ".ident directive not supported");
   OS << "\t.ident\t";