[TTI] Make the cost APIs in TargetTransformInfo consistently use 'int'

[oota-llvm.git] / lib / Target / X86 / X86TargetMachine.cpp

diff --git a/lib/Target/X86/X86TargetMachine.cpp b/lib/Target/X86/X86TargetMachine.cpp
index f56d6bab157027acee43d71a3ecfcc4a38c0ffda..fb9cb4ba4c86edf8c0c9158fab35d1935b38265e 100644 (file)
--- a/lib/Target/X86/X86TargetMachine.cpp
+++ b/lib/Target/X86/X86TargetMachine.cpp
@@ -24,6 +24,10 @@
 #include "llvm/Target/TargetOptions.h"
 using namespace llvm;
 
+static cl::opt<bool> EnableMachineCombinerPass("x86-machine-combiner",
+                               cl::desc("Enable the machine combiner pass"),
+                               cl::init(true), cl::Hidden);
+
 extern "C" void LLVMInitializeX86Target() {
   // Register the target.
   RegisterTargetMachine<X86TargetMachine> X(TheX86_32Target);
@@ -37,12 +41,10 @@ static std::unique_ptr<TargetLoweringObjectFile> createTLOF(const Triple &TT) {
     return make_unique<TargetLoweringObjectFileMachO>();
   }
 
-  if (TT.isPS4CPU())
-    return make_unique<PS4TargetObjectFile>();
-  if (TT.isOSLinux())
-    return make_unique<X86LinuxTargetObjectFile>();
+  if (TT.isOSLinux() || TT.isOSNaCl())
+    return make_unique<X86LinuxNaClTargetObjectFile>();
   if (TT.isOSBinFormatELF())
-    return make_unique<TargetLoweringObjectFileELF>();
+    return make_unique<X86ELFTargetObjectFile>();
   if (TT.isKnownWindowsMSVCEnvironment())
     return make_unique<X86WindowsTargetObjectFile>();
   if (TT.isOSBinFormatCOFF())
@@ -83,7 +85,7 @@ static std::string computeDataLayout(const Triple &TT) {
 
   // The stack is aligned to 32 bits on some ABIs and 128 bits on others.
   if (!TT.isArch64Bit() && TT.isOSWindows())
-    Ret += "-S32";
+    Ret += "-a:0:32-S32";
   else
     Ret += "-S128";
 
@@ -92,18 +94,15 @@ static std::string computeDataLayout(const Triple &TT) {
 
 /// X86TargetMachine ctor - Create an X86 target.
 ///
-X86TargetMachine::X86TargetMachine(const Target &T, StringRef TT, StringRef CPU,
-                                   StringRef FS, const TargetOptions &Options,
+X86TargetMachine::X86TargetMachine(const Target &T, const Triple &TT,
+                                   StringRef CPU, StringRef FS,
+                                   const TargetOptions &Options,
                                    Reloc::Model RM, CodeModel::Model CM,
                                    CodeGenOpt::Level OL)
-    : LLVMTargetMachine(T, TT, CPU, FS, Options, RM, CM, OL),
-      TLOF(createTLOF(Triple(getTargetTriple()))),
-      DL(computeDataLayout(Triple(TT))),
+    : LLVMTargetMachine(T, computeDataLayout(TT), TT, CPU, FS, Options, RM, CM,
+                        OL),
+      TLOF(createTLOF(getTargetTriple())),
       Subtarget(TT, CPU, FS, *this, Options.StackAlignmentOverride) {
-  // default to hard float ABI
-  if (Options.FloatABIType == FloatABI::Default)
-    this->Options.FloatABIType = FloatABI::Hard;
-
   // Windows stack unwinder gets confused when execution flow "falls through"
   // after a call to 'noreturn' function.
   // To prevent that, we emit a trap for 'unreachable' IR instructions.
@@ -111,6 +110,16 @@ X86TargetMachine::X86TargetMachine(const Target &T, StringRef TT, StringRef CPU,
   if (Subtarget.isTargetWin64())
     this->Options.TrapUnreachable = true;
 
+  // By default (and when -ffast-math is on), enable estimate codegen for
+  // everything except scalar division. By default, use 1 refinement step for
+  // all operations. Defaults may be overridden by using command-line options.
+  // Scalar division estimates are disabled because they break too much
+  // real-world code. These defaults match GCC behavior.
+  this->Options.Reciprocals.setDefaults("sqrtf", true, 1);
+  this->Options.Reciprocals.setDefaults("divf", false, 1);
+  this->Options.Reciprocals.setDefaults("vec-sqrtf", true, 1);
+  this->Options.Reciprocals.setDefaults("vec-divf", true, 1);
+
   initAsmInfo();
 }
 
@@ -133,13 +142,15 @@ X86TargetMachine::getSubtargetImpl(const Function &F) const {
   // function before we can generate a subtarget. We also need to use
   // it as a key for the subtarget since that can be the only difference
   // between two functions.
-  Attribute SFAttr = F.getFnAttribute("use-soft-float");
-  bool SoftFloat = !SFAttr.hasAttribute(Attribute::None)
-                       ? SFAttr.getValueAsString() == "true"
-                       : Options.UseSoftFloat;
-
-  auto &I = SubtargetMap[CPU + FS + (SoftFloat ? "use-soft-float=true"
-                                               : "use-soft-float=false")];
+  bool SoftFloat =
+      F.hasFnAttribute("use-soft-float") &&
+      F.getFnAttribute("use-soft-float").getValueAsString() == "true";
+  // If the soft float attribute is set on the function turn on the soft float
+  // subtarget feature.
+  if (SoftFloat)
+    FS += FS.empty() ? "+soft-float" : ",+soft-float";
+
+  auto &I = SubtargetMap[CPU + FS];
   if (!I) {
     // This needs to be done before we create a new subtarget since any
     // creation will depend on the TM and the code generation flags on the
@@ -187,9 +198,11 @@ public:
   void addIRPasses() override;
   bool addInstSelector() override;
   bool addILPOpts() override;
+  bool addPreISel() override;
   void addPreRegAlloc() override;
   void addPostRegAlloc() override;
   void addPreEmitPass() override;
+  void addPreSched2() override;
 };
 } // namespace
 
@@ -208,7 +221,7 @@ bool X86PassConfig::addInstSelector() {
   addPass(createX86ISelDag(getX86TargetMachine(), getOptLevel()));
 
   // For ELF, cleanup any local-dynamic TLS accesses.
-  if (Triple(TM->getTargetTriple()).isOSBinFormatELF() &&
+  if (TM->getTargetTriple().isOSBinFormatELF() &&
       getOptLevel() != CodeGenOpt::None)
     addPass(createCleanupLocalDynamicTLSPass());
 
@@ -219,6 +232,16 @@ bool X86PassConfig::addInstSelector() {
 
 bool X86PassConfig::addILPOpts() {
   addPass(&EarlyIfConverterID);
+  if (EnableMachineCombinerPass)
+    addPass(&MachineCombinerID);
+  return true;
+}
+
+bool X86PassConfig::addPreISel() {
+  // Only add this pass for 32-bit x86 Windows.
+  const Triple &TT = TM->getTargetTriple();
+  if (TT.isOSWindows() && TT.getArch() == Triple::x86)
+    addPass(createX86WinEHStatePass());
   return true;
 }
 
@@ -230,6 +253,8 @@ void X86PassConfig::addPostRegAlloc() {
   addPass(createX86FloatingPointStackifierPass());
 }
 
+void X86PassConfig::addPreSched2() { addPass(createX86ExpandPseudoPass()); }
+
 void X86PassConfig::addPreEmitPass() {
   if (getOptLevel() != CodeGenOpt::None)
     addPass(createExecutionDependencyFixPass(&X86::VR128RegClass));