COFF: Implement parallel LTO code generation.

pcc · pcc · commit df5783b7a5a3 · 2015-08-28T22:16:09.000Z
This is exposed via a new flag /opt:lldltojobs=N, where N is the number of code generation threads. Differential Revision: http://reviews.llvm.org/D12309 llvm-svn: 246342
diff --git a/lld/COFF/Config.h b/lld/COFF/Config.h
@@ -91,6 +91,9 @@ struct Configuration {
   // Used for /opt:lldlto=N
   unsigned LTOOptLevel = 2;
 
+  // Used for /opt:lldltojobs=N
+  unsigned LTOJobs = 1;
+
   // Used for /merge:from=to (e.g. /merge:.rdata=.text)
   std::map<StringRef, StringRef> Merge;
 
diff --git a/lld/COFF/Driver.cpp b/lld/COFF/Driver.cpp
@@ -372,9 +372,14 @@ void LinkerDriver::link(llvm::ArrayRef<const char *> ArgsArr) {
     if (StringRef(S).startswith("lldlto=")) {
       StringRef OptLevel = StringRef(S).substr(7);
       if (OptLevel.getAsInteger(10, Config->LTOOptLevel) ||
-          Config->LTOOptLevel > 3) {
+          Config->LTOOptLevel > 3)
         error("/opt:lldlto: invalid optimization level: " + OptLevel);
-      }
+      continue;
+    }
+    if (StringRef(S).startswith("lldltojobs=")) {
+      StringRef Jobs = StringRef(S).substr(11);
+      if (Jobs.getAsInteger(10, Config->LTOJobs) || Config->LTOJobs == 0)
+        error("/opt:lldltojobs: invalid job count: " + Jobs);
       continue;
     }
     if (S != "ref" && S != "icf" && S != "noicf" &&
@@ -594,9 +599,9 @@ void LinkerDriver::link(llvm::ArrayRef<const char *> ArgsArr) {
     Symtab.run();
   }
 
-  // Do LTO by compiling bitcode input files to a native COFF file
-  // then link that file.
-  Symtab.addCombinedLTOObject();
+  // Do LTO by compiling bitcode input files to a set of native COFF files then
+  // link those files.
+  Symtab.addCombinedLTOObjects();
 
   // Make sure we have resolved all symbols.
   Symtab.reportRemainingUndefines(/*Resolve=*/true);
diff --git a/lld/COFF/SymbolTable.cpp b/lld/COFF/SymbolTable.cpp
@@ -331,21 +331,7 @@ void SymbolTable::printMap(llvm::raw_ostream &OS) {
   }
 }
 
-void SymbolTable::addCombinedLTOObject() {
-  if (BitcodeFiles.empty())
-    return;
-
-  // Diagnose any undefined symbols early, but do not resolve weak externals,
-  // as resolution breaks the invariant that each Symbol points to a unique
-  // SymbolBody, which we rely on to replace DefinedBitcode symbols correctly.
-  reportRemainingUndefines(/*Resolve=*/false);
-
-  // Create an object file and add it to the symbol table by replacing any
-  // DefinedBitcode symbols with the definitions in the object file.
-  LTOCodeGenerator CG;
-  CG.setOptLevel(Config->LTOOptLevel);
-  ObjectFile *Obj = createLTOObject(&CG);
-
+void SymbolTable::addCombinedLTOObject(ObjectFile *Obj) {
   for (SymbolBody *Body : Obj->getSymbols()) {
     if (!Body->isExternal())
       continue;
@@ -371,6 +357,25 @@ void SymbolTable::addCombinedLTOObject() {
     if (Comp < 0)
       Sym->Body = Body;
   }
+}
+
+void SymbolTable::addCombinedLTOObjects() {
+  if (BitcodeFiles.empty())
+    return;
+
+  // Diagnose any undefined symbols early, but do not resolve weak externals,
+  // as resolution breaks the invariant that each Symbol points to a unique
+  // SymbolBody, which we rely on to replace DefinedBitcode symbols correctly.
+  reportRemainingUndefines(/*Resolve=*/false);
+
+  // Create an object file and add it to the symbol table by replacing any
+  // DefinedBitcode symbols with the definitions in the object file.
+  LTOCodeGenerator CG;
+  CG.setOptLevel(Config->LTOOptLevel);
+  std::vector<ObjectFile *> Objs = createLTOObjects(&CG);
+
+  for (ObjectFile *Obj : Objs)
+    addCombinedLTOObject(Obj);
 
   size_t NumBitcodeFiles = BitcodeFiles.size();
   run();
@@ -379,8 +384,8 @@ void SymbolTable::addCombinedLTOObject() {
 }
 
 // Combine and compile bitcode files and then return the result
-// as a regular COFF object file.
-ObjectFile *SymbolTable::createLTOObject(LTOCodeGenerator *CG) {
+// as a vector of regular COFF object files.
+std::vector<ObjectFile *> SymbolTable::createLTOObjects(LTOCodeGenerator *CG) {
   // All symbols referenced by non-bitcode objects must be preserved.
   for (ObjectFile *File : ObjectFiles)
     for (SymbolBody *Body : File->getSymbols())
@@ -406,14 +411,32 @@ ObjectFile *SymbolTable::createLTOObject(LTOCodeGenerator *CG) {
     CG->addModule(BitcodeFiles[I]->getModule());
 
   std::string ErrMsg;
-  LTOMB = CG->compile(false, false, false, ErrMsg); // take MB ownership
-  if (!LTOMB)
+  if (!CG->optimize(false, false, false, ErrMsg))
+    error(ErrMsg);
+
+  Objs.resize(Config->LTOJobs);
+  // Use std::list to avoid invalidation of pointers in OSPtrs.
+  std::list<raw_svector_ostream> OSs;
+  std::vector<raw_pwrite_stream *> OSPtrs;
+  for (SmallVector<char, 0> &Obj : Objs) {
+    OSs.emplace_back(Obj);
+    OSPtrs.push_back(&OSs.back());
+  }
+
+  if (!CG->compileOptimized(OSPtrs, ErrMsg))
     error(ErrMsg);
-  auto *Obj = new ObjectFile(LTOMB->getMemBufferRef());
-  Files.emplace_back(Obj);
-  ObjectFiles.push_back(Obj);
-  Obj->parse();
-  return Obj;
+
+  std::vector<ObjectFile *> ObjFiles;
+  for (SmallVector<char, 0> &Obj : Objs) {
+    auto *ObjFile = new ObjectFile(
+        MemoryBufferRef(StringRef(Obj.data(), Obj.size()), "<LTO object>"));
+    Files.emplace_back(ObjFile);
+    ObjectFiles.push_back(ObjFile);
+    ObjFile->parse();
+    ObjFiles.push_back(ObjFile);
+  }
+
+  return ObjFiles;
 }
 
 } // namespace coff
diff --git a/lld/COFF/SymbolTable.h b/lld/COFF/SymbolTable.h
@@ -68,10 +68,10 @@ class SymbolTable {
   // Print a layout map to OS.
   void printMap(llvm::raw_ostream &OS);
 
-  // Build a COFF object representing the combined contents of BitcodeFiles
-  // and add it to the symbol table. Called after all files are added and
-  // before the writer writes results to a file.
-  void addCombinedLTOObject();
+  // Build a set of COFF objects representing the combined contents of
+  // BitcodeFiles and add them to the symbol table. Called after all files are
+  // added and before the writer writes results to a file.
+  void addCombinedLTOObjects();
 
   // The writer needs to handle DLL import libraries specially in
   // order to create the import descriptor table.
@@ -98,7 +98,8 @@ class SymbolTable {
   StringRef findByPrefix(StringRef Prefix);
 
   void addMemberFile(Lazy *Body);
-  ObjectFile *createLTOObject(llvm::LTOCodeGenerator *CG);
+  void addCombinedLTOObject(ObjectFile *Obj);
+  std::vector<ObjectFile *> createLTOObjects(llvm::LTOCodeGenerator *CG);
 
   llvm::DenseMap<StringRef, Symbol *> Symtab;
 
@@ -107,7 +108,7 @@ class SymbolTable {
   std::vector<InputFile *> ObjectQueue;
 
   std::vector<BitcodeFile *> BitcodeFiles;
-  std::unique_ptr<MemoryBuffer> LTOMB;
+  std::vector<SmallVector<char, 0>> Objs;
   llvm::BumpPtrAllocator Alloc;
 };
 
diff --git a/lld/test/COFF/lto-parallel.ll b/lld/test/COFF/lto-parallel.ll
@@ -0,0 +1,20 @@
+; RUN: llvm-as -o %t.obj %s
+; RUN: lld-link /out:%t.exe /entry:foo /include:bar /opt:lldltojobs=2 /subsystem:console /lldmap:%t.map %t.obj
+; RUN: FileCheck %s < %t.map
+
+target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc"
+
+; CHECK: <lto object>:
+; CHECK: foo
+define void @foo() {
+  call void @bar()
+  ret void
+}
+
+; CHECK: <lto object>:
+; CHECK: bar
+define void @bar() {
+  call void @foo()
+  ret void
+}
diff --git a/lld/test/COFF/weak-external.test b/lld/test/COFF/weak-external.test
@@ -4,7 +4,7 @@
 # RUN: lld-link /out:%t2.exe /entry:g /subsystem:console /lldmap:%t2.map %t.obj %t.lto.obj
 # RUN: FileCheck %s < %t2.map
 
-# CHECK: lto-llvm{{.*}}:
+# CHECK: <lto object>:
 # CHECK-NOT: :
 # CHECK: {{ g$}}
 
diff --git a/lld/test/COFF/weak-external3.test b/lld/test/COFF/weak-external3.test
@@ -5,7 +5,7 @@
 # RUN: lld-link /out:%t2.exe /entry:f /subsystem:console /lldmap:%t2.map %t.obj %t.lto.obj
 # RUN: FileCheck --check-prefix=CHECK2 %s < %t2.map
 
-# CHECK1: lto-llvm{{.*}}:
+# CHECK1: <lto object>:
 # CHECK1-NOT: :
 # CHECK1: {{ g$}}
 
