s/DebugInfoEnumerator/DebugInfoFinder/g

[oota-llvm.git] / lib / Transforms / Scalar / MemCpyOptimizer.cpp

diff --git a/lib/Transforms/Scalar/MemCpyOptimizer.cpp b/lib/Transforms/Scalar/MemCpyOptimizer.cpp
index 5d68388a99c0df766f3d896fa7a19aac21b5fcc9..224a13664832a94075a31d96c8ebf28b607b1a07 100644 (file)
--- a/lib/Transforms/Scalar/MemCpyOptimizer.cpp
+++ b/lib/Transforms/Scalar/MemCpyOptimizer.cpp
@@ -16,6 +16,7 @@
 #include "llvm/Transforms/Scalar.h"
 #include "llvm/IntrinsicInst.h"
 #include "llvm/Instructions.h"
+#include "llvm/LLVMContext.h"
 #include "llvm/ADT/SmallVector.h"
 #include "llvm/ADT/Statistic.h"
 #include "llvm/Analysis/Dominators.h"
@@ -35,7 +36,7 @@ STATISTIC(NumMemSetInfer, "Number of memsets inferred");
 /// true for all i8 values obviously, but is also true for i32 0, i32 -1,
 /// i16 0xF0F0, double 0.0 etc.  If the value can't be handled with a repeated
 /// byte store (e.g. i16 0x1234), return null.
-static Value *isBytewiseValue(Value *V) {
+static Value *isBytewiseValue(Value *V, LLVMContext& Context) {
   // All byte-wide stores are splatable, even of arbitrary variables.
   if (V->getType() == Type::Int8Ty) return V;
   
@@ -68,7 +69,7 @@ static Value *isBytewiseValue(Value *V) {
         if (Val != Val2)
           return 0;
       }
-      return ConstantInt::get(Val);
+      return ConstantInt::get(Context, Val);
     }
   }
   
@@ -104,7 +105,7 @@ static int64_t GetOffsetFromIndex(const GetElementPtrInst *GEP, unsigned Idx,
     
     // Otherwise, we have a sequential type like an array or vector.  Multiply
     // the index by the ElementSize.
-    uint64_t Size = TD.getABITypeSize(GTI.getIndexedType());
+    uint64_t Size = TD.getTypeAllocSize(GTI.getIndexedType());
     Offset += Size*OpC->getSExtValue();
   }
 
@@ -277,7 +278,7 @@ void MemsetRanges::addStore(int64_t Start, StoreInst *SI) {
   // End.
   if (End > I->End) {
     I->End = End;
-    range_iterator NextI = I;;
+    range_iterator NextI = I;
     while (++NextI != E && End >= NextI->Start) {
       // Merge the range in.
       I->TheStores.append(NextI->TheStores.begin(), NextI->TheStores.end());
@@ -345,12 +346,13 @@ bool MemCpyOpt::processStore(StoreInst *SI, BasicBlock::iterator& BBI) {
   // Ensure that the value being stored is something that can be memset'able a
   // byte at a time like "0" or "-1" or any width, as well as things like
   // 0xA0A0A0A0 and 0.0.
-  Value *ByteVal = isBytewiseValue(SI->getOperand(0));
+  Value *ByteVal = isBytewiseValue(SI->getOperand(0), SI->getContext());
   if (!ByteVal)
     return false;
 
   TargetData &TD = getAnalysis<TargetData>();
   AliasAnalysis &AA = getAnalysis<AliasAnalysis>();
+  Module *M = SI->getParent()->getParent()->getParent();
 
   // Okay, so we now have a single store that can be splatable.  Scan to find
   // all subsequent stores of the same value to offset from the same pointer.
@@ -384,7 +386,8 @@ bool MemCpyOpt::processStore(StoreInst *SI, BasicBlock::iterator& BBI) {
     if (NextStore->isVolatile()) break;
     
     // Check to see if this stored value is of the same byte-splattable value.
-    if (ByteVal != isBytewiseValue(NextStore->getOperand(0)))
+    if (ByteVal != isBytewiseValue(NextStore->getOperand(0), 
+                                   NextStore->getContext()))
       break;
 
     // Check to see if this store is to a constant offset from the start ptr.
@@ -429,8 +432,7 @@ bool MemCpyOpt::processStore(StoreInst *SI, BasicBlock::iterator& BBI) {
   
     if (MemSetF == 0) {
       const Type *Tys[] = {Type::Int64Ty};
-      MemSetF = Intrinsic::getDeclaration(SI->getParent()->getParent()
-                                          ->getParent(), Intrinsic::memset,
+      MemSetF = Intrinsic::getDeclaration(M, Intrinsic::memset,
                                           Tys, 1);
    }
     
@@ -440,13 +442,15 @@ bool MemCpyOpt::processStore(StoreInst *SI, BasicBlock::iterator& BBI) {
     // Cast the start ptr to be i8* as memset requires.
     const Type *i8Ptr = PointerType::getUnqual(Type::Int8Ty);
     if (StartPtr->getType() != i8Ptr)
-      StartPtr = new BitCastInst(StartPtr, i8Ptr, StartPtr->getNameStart(),
+      StartPtr = new BitCastInst(StartPtr, i8Ptr, StartPtr->getName(),
                                  InsertPt);
   
     Value *Ops[] = {
       StartPtr, ByteVal,   // Start, value
-      ConstantInt::get(Type::Int64Ty, Range.End-Range.Start),  // size
-      ConstantInt::get(Type::Int32Ty, Range.Alignment)   // align
+      // size
+      ConstantInt::get(Type::Int64Ty, Range.End-Range.Start),
+      // align
+      ConstantInt::get(Type::Int32Ty, Range.Alignment)
     };
     Value *C = CallInst::Create(MemSetF, Ops, Ops+4, "", InsertPt);
     DEBUG(cerr << "Replace stores:\n";
@@ -511,7 +515,7 @@ bool MemCpyOpt::performCallSlotOptzn(MemCpyInst *cpy, CallInst *C) {
   if (!srcArraySize)
     return false;
 
-  uint64_t srcSize = TD.getABITypeSize(srcAlloca->getAllocatedType()) *
+  uint64_t srcSize = TD.getTypeAllocSize(srcAlloca->getAllocatedType()) *
     srcArraySize->getZExtValue();
 
   if (cpyLength->getZExtValue() < srcSize)
@@ -526,7 +530,7 @@ bool MemCpyOpt::performCallSlotOptzn(MemCpyInst *cpy, CallInst *C) {
     if (!destArraySize)
       return false;
 
-    uint64_t destSize = TD.getABITypeSize(A->getAllocatedType()) *
+    uint64_t destSize = TD.getTypeAllocSize(A->getAllocatedType()) *
       destArraySize->getZExtValue();
 
     if (destSize < srcSize)
@@ -538,7 +542,7 @@ bool MemCpyOpt::performCallSlotOptzn(MemCpyInst *cpy, CallInst *C) {
       return false;
 
     const Type* StructTy = cast<PointerType>(A->getType())->getElementType();
-    uint64_t destSize = TD.getABITypeSize(StructTy);
+    uint64_t destSize = TD.getTypeAllocSize(StructTy);
 
     if (destSize < srcSize)
       return false;
@@ -630,13 +634,12 @@ bool MemCpyOpt::processMemCpy(MemCpyInst* M) {
   //   a) memcpy-memcpy xform which exposes redundance for DSE
   //   b) call-memcpy xform for return slot optimization
   MemDepResult dep = MD.getDependency(M);
-  if (!dep.isNormal())
+  if (!dep.isClobber())
     return false;
-  else if (!isa<MemCpyInst>(dep.getInst())) {
+  if (!isa<MemCpyInst>(dep.getInst())) {
     if (CallInst* C = dyn_cast<CallInst>(dep.getInst()))
       return performCallSlotOptzn(M, C);
-    else
-      return false;
+    return false;
   }
   
   MemCpyInst* MDep = cast<MemCpyInst>(dep.getInst());
@@ -679,13 +682,11 @@ bool MemCpyOpt::processMemCpy(MemCpyInst* M) {
                                  M->getParent()->getParent()->getParent(),
                                  M->getIntrinsicID(), Tys, 1);
     
-  std::vector<Value*> args;
-  args.push_back(M->getRawDest());
-  args.push_back(MDep->getRawSource());
-  args.push_back(M->getLength());
-  args.push_back(M->getAlignment());
+  Value *Args[4] = {
+    M->getRawDest(), MDep->getRawSource(), M->getLength(), M->getAlignmentCst()
+  };
   
-  CallInst* C = CallInst::Create(MemCpyFun, args.begin(), args.end(), "", M);
+  CallInst* C = CallInst::Create(MemCpyFun, Args, Args+4, "", M);
   
   
   // If C and M don't interfere, then this is a valid transformation.  If they