binary serialization prototype

Author: isuckatcs

clang/include/clang/Driver/Options.td

@@ -5956,7 +5956,7 @@ def emit_summaries : Flag<["-", "--"], "emit-summaries">, Flags<[NoXarchOption]>
   HelpText<"Alias for --emit-summaries=cwd">;
 def summary_format_EQ : Joined<["-", "--"], "summary-format=">, Flags<[NoXarchOption]>,
   Visibility<[ClangOption, CC1Option]>,
-  HelpText<"The format of the emitted summaries. Can be set to 'json' (default) or 'yaml'">;
+  HelpText<"The format of the emitted summaries. Can be set to 'json' (default), 'yaml', or 'binary'">;
 def save_stats_EQ : Joined<["-", "--"], "save-stats=">, Flags<[NoXarchOption]>,
   HelpText<"Save llvm statistics.">;
 def save_stats : Flag<["-", "--"], "save-stats">, Flags<[NoXarchOption]>,

clang/include/clang/Summary/SummarySerialization.h

@@ -2,6 +2,8 @@
 #define LLVM_CLANG_SUMMARY_SUMMARYSERIALIZATION_H
 
 #include "clang/Summary/SummaryContext.h"
+#include "llvm/Bitstream/BitstreamReader.h"
+#include "llvm/Bitstream/BitstreamWriter.h"
 
 namespace clang {
 class SummarySerializer {
@@ -38,6 +40,63 @@ class YAMLSummarySerializer : public SummarySerializer {
                  raw_ostream &OS) override;
   void parse(StringRef) override;
 };
+
+class BinarySummarySerializer : public SummarySerializer {
+  enum BlockIDs {
+    ATTRIBUTE_BLOCK_ID = llvm::bitc::FIRST_APPLICATION_BLOCKID,
+    IDENTIFIER_BLOCK_ID,
+    SUMMARY_BLOCK_ID
+  };
+
+  enum AttributeRecordTypes {
+    ATTR = 1,
+  };
+
+  enum IdentifierRecordTypes {
+    IDENTIFIER = 1,
+  };
+
+  enum SummaryRecordTypes { FUNCTION = 1 };
+
+  // FIXME: get rid of this global state
+  std::map<const SummaryAttr *, uint64_t> AttrIDs;
+  std::map<std::string, uint64_t> FunctionIDs;
+
+  std::vector<const SummaryAttr *> ParsedAttrIDs;
+  std::vector<std::string> ParsedFunctionIDs;
+
+  llvm::SmallVector<char, 32> Buffer;
+  llvm::BitstreamWriter Stream;
+
+  void PopulateBlockInfo();
+  void EmitAttributeBlock();
+  void EmitIdentifierBlock();
+  void EmitSummaryBlock();
+
+  void EmitBlock(unsigned ID, const char *Name);
+  void EmitRecord(unsigned ID, const char *Name);
+
+  llvm::Error handleBlockStartCommon(unsigned ID,
+                                     llvm::BitstreamCursor &Stream);
+  llvm::Error handleBlockRecordsCommon(
+      llvm::BitstreamCursor &Stream,
+      llvm::function_ref<void(const SmallVector<uint64_t, 64> &)>);
+
+  llvm::Error parseAttributeBlock(llvm::BitstreamCursor &Stream);
+  llvm::Error parseIdentifierBlock(llvm::BitstreamCursor &Stream);
+  llvm::Error parseSummaryBlock(llvm::BitstreamCursor &Stream);
+  llvm::Error parseBlock(unsigned ID, llvm::BitstreamCursor &Stream);
+  llvm::Error parseImpl(StringRef Buffer);
+
+public:
+  BinarySummarySerializer(SummaryContext &SummaryCtx)
+      : SummarySerializer(SummaryCtx), Stream(Buffer) {};
+
+  void serialize(const std::vector<std::unique_ptr<FunctionSummary>> &,
+                 raw_ostream &OS) override;
+  void parse(StringRef) override;
+};
+
 } // namespace clang
 
 #endif // LLVM_CLANG_SUMMARY_SUMMARYSERIALIZATION_H

clang/lib/Frontend/CompilerInstance.cpp

@@ -750,8 +750,10 @@ void CompilerInstance::createSummaryConsumer(FrontendInputFile Input) {
 
   llvm::SmallString<32> SummaryFile = EmitSummaryDir;
   llvm::sys::path::append(SummaryFile, Input.getFile());
+
+  StringRef Format = getFrontendOpts().SummaryFormat;
   llvm::sys::path::replace_extension(SummaryFile,
-                                     getFrontendOpts().SummaryFormat);
+                                     Format == "binary" ? "summary" : Format);
 
   std::error_code EC;
   SummaryOS.reset(new llvm::raw_fd_ostream(SummaryFile, EC,
@@ -778,6 +780,8 @@ void CompilerInstance::createSummarySerializer() {
 
   if (Format == "yaml")
     Serializer = new YAMLSummarySerializer(*getSummaryContext());
+  else if (Format == "binary")
+    Serializer = new BinarySummarySerializer(*getSummaryContext());
   else
     Serializer = new JSONSummarySerializer(*getSummaryContext());
 

clang/lib/Frontend/CompilerInvocation.cpp

@@ -3271,7 +3271,7 @@ static bool ParseFrontendArgs(FrontendOptions &Opts, ArgList &Args,
     StringRef Format = A->getValue();
 
     // FIXME: don't hardcode these values
-    if (Format == "yaml")
+    if (Format == "yaml" || Format == "binary")
       Opts.SummaryFormat = Format;
   };
 

clang/lib/Frontend/FrontendAction.cpp

@@ -986,7 +986,9 @@ bool FrontendAction::BeginSourceFile(CompilerInstance &CI,
     static const char *reducedCacheName =
         "reduced-summary-so-that-we-do-not-have-to-evaluate-it-every-time";
     const std::string summaryExtension =
-        '.' + CI.getFrontendOpts().SummaryFormat;
+        '.' + (CI.getFrontendOpts().SummaryFormat == "binary"
+                   ? "summary"
+                   : CI.getFrontendOpts().SummaryFormat);
 
     FileManager &FileMgr = CI.getFileManager();
     StringRef SummaryDirPath = CI.getFrontendOpts().ReadSummaryDir;