Lift the speculation visitor above all the helpers that are targeted at

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

diff --git a/lib/Transforms/Scalar/MemCpyOptimizer.cpp b/lib/Transforms/Scalar/MemCpyOptimizer.cpp
index 2a5ee33eb1ed7b630a5017a1ff58002b9422ffd0..c02d4cec4cf74b0a05b9c52f2ada54b78a541b4d 100644 (file)
--- a/lib/Transforms/Scalar/MemCpyOptimizer.cpp
+++ b/lib/Transforms/Scalar/MemCpyOptimizer.cpp
@@ -38,7 +38,7 @@ STATISTIC(NumMemSetInfer, "Number of memsets inferred");
 STATISTIC(NumMoveToCpy,   "Number of memmoves converted to memcpy");
 STATISTIC(NumCpyToSet,    "Number of memcpys converted to memset");
 
-static int64_t GetOffsetFromIndex(const GetElementPtrInst *GEP, unsigned Idx,
+static int64_t GetOffsetFromIndex(const GEPOperator *GEP, unsigned Idx,
                                   bool &VariableIdxFound, const TargetData &TD){
   // Skip over the first indices.
   gep_type_iterator GTI = gep_type_begin(GEP);
@@ -75,8 +75,8 @@ static bool IsPointerOffset(Value *Ptr1, Value *Ptr2, int64_t &Offset,
                             const TargetData &TD) {
   Ptr1 = Ptr1->stripPointerCasts();
   Ptr2 = Ptr2->stripPointerCasts();
-  GetElementPtrInst *GEP1 = dyn_cast<GetElementPtrInst>(Ptr1);
-  GetElementPtrInst *GEP2 = dyn_cast<GetElementPtrInst>(Ptr2);
+  GEPOperator *GEP1 = dyn_cast<GEPOperator>(Ptr1);
+  GEPOperator *GEP2 = dyn_cast<GEPOperator>(Ptr2);
 
   bool VariableIdxFound = false;
 
@@ -332,7 +332,7 @@ namespace {
     bool processMemCpy(MemCpyInst *M);
     bool processMemMove(MemMoveInst *M);
     bool performCallSlotOptzn(Instruction *cpy, Value *cpyDst, Value *cpySrc,
-                              uint64_t cpyLen, CallInst *C);
+                              uint64_t cpyLen, unsigned cpyAlign, CallInst *C);
     bool processMemCpyMemCpyDependence(MemCpyInst *M, MemCpyInst *MDep,
                                        uint64_t MSize);
     bool processByValArgument(CallSite CS, unsigned ArgNo);
@@ -509,10 +509,18 @@ bool MemCpyOpt::processStore(StoreInst *SI, BasicBlock::iterator &BBI) {
       }
 
       if (C) {
+        unsigned storeAlign = SI->getAlignment();
+        if (!storeAlign)
+          storeAlign = TD->getABITypeAlignment(SI->getOperand(0)->getType());
+        unsigned loadAlign = LI->getAlignment();
+        if (!loadAlign)
+          loadAlign = TD->getABITypeAlignment(LI->getType());
+
         bool changed = performCallSlotOptzn(LI,
                         SI->getPointerOperand()->stripPointerCasts(),
                         LI->getPointerOperand()->stripPointerCasts(),
-                        TD->getTypeStoreSize(SI->getOperand(0)->getType()), C);
+                        TD->getTypeStoreSize(SI->getOperand(0)->getType()),
+                        std::min(storeAlign, loadAlign), C);
         if (changed) {
           MD->removeInstruction(SI);
           SI->eraseFromParent();
@@ -559,7 +567,8 @@ bool MemCpyOpt::processMemSet(MemSetInst *MSI, BasicBlock::iterator &BBI) {
 /// the call write its result directly into the destination of the memcpy.
 bool MemCpyOpt::performCallSlotOptzn(Instruction *cpy,
                                      Value *cpyDest, Value *cpySrc,
-                                     uint64_t cpyLen, CallInst *C) {
+                                     uint64_t cpyLen, unsigned cpyAlign,
+                                     CallInst *C) {
   // The general transformation to keep in mind is
   //
   //   call @func(..., src, ...)
@@ -596,6 +605,16 @@ bool MemCpyOpt::performCallSlotOptzn(Instruction *cpy,
   if (cpyLen < srcSize)
     return false;
 
+  // Check that dest points to memory that is at least as aligned as src.
+  unsigned srcAlign = srcAlloca->getAlignment();
+  if (!srcAlign)
+    srcAlign = TD->getABITypeAlignment(srcAlloca->getAllocatedType());
+  bool isDestSufficientlyAligned = srcAlign <= cpyAlign;
+  // If dest is not aligned enough and we can't increase its alignment then
+  // bail out.
+  if (!isDestSufficientlyAligned && !isa<AllocaInst>(cpyDest))
+    return false;
+
   // Check that accessing the first srcSize bytes of dest will not cause a
   // trap.  Otherwise the transform is invalid since it might cause a trap
   // to occur earlier than it otherwise would.
@@ -673,20 +692,26 @@ bool MemCpyOpt::performCallSlotOptzn(Instruction *cpy,
   bool changedArgument = false;
   for (unsigned i = 0; i < CS.arg_size(); ++i)
     if (CS.getArgument(i)->stripPointerCasts() == cpySrc) {
-      if (cpySrc->getType() != cpyDest->getType())
-        cpyDest = CastInst::CreatePointerCast(cpyDest, cpySrc->getType(),
-                                              cpyDest->getName(), C);
+      Value *Dest = cpySrc->getType() == cpyDest->getType() ?  cpyDest
+        : CastInst::CreatePointerCast(cpyDest, cpySrc->getType(),
+                                      cpyDest->getName(), C);
       changedArgument = true;
-      if (CS.getArgument(i)->getType() == cpyDest->getType())
-        CS.setArgument(i, cpyDest);
+      if (CS.getArgument(i)->getType() == Dest->getType())
+        CS.setArgument(i, Dest);
       else
-        CS.setArgument(i, CastInst::CreatePointerCast(cpyDest,
-                          CS.getArgument(i)->getType(), cpyDest->getName(), C));
+        CS.setArgument(i, CastInst::CreatePointerCast(Dest,
+                          CS.getArgument(i)->getType(), Dest->getName(), C));
     }
 
   if (!changedArgument)
     return false;
 
+  // If the destination wasn't sufficiently aligned then increase its alignment.
+  if (!isDestSufficientlyAligned) {
+    assert(isa<AllocaInst>(cpyDest) && "Can only increase alloca alignment!");
+    cast<AllocaInst>(cpyDest)->setAlignment(srcAlign);
+  }
+
   // Drop any cached information about the call, because we may have changed
   // its dependence information by changing its parameter.
   MD->removeInstruction(C);
@@ -813,7 +838,8 @@ bool MemCpyOpt::processMemCpy(MemCpyInst *M) {
   if (DepInfo.isClobber()) {
     if (CallInst *C = dyn_cast<CallInst>(DepInfo.getInst())) {
       if (performCallSlotOptzn(M, M->getDest(), M->getSource(),
-                               CopySize->getZExtValue(), C)) {
+                               CopySize->getZExtValue(), M->getAlignment(),
+                               C)) {
         MD->removeInstruction(M);
         M->eraseFromParent();
         return true;