Revert "Revert "Mark vastart_save_xmm_regs as changing EFLAGS""

[oota-llvm.git] / lib / Target / X86 / X86ISelLowering.cpp

diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index e87e7edab3c83253f17da7d42dd94585fef2b806..c780c14abbc3c19b8fa74c43da5767698af22027 100644 (file)
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -1356,7 +1356,7 @@ void X86TargetLowering::resetOperationActions() {
     setOperationAction(ISD::FP_ROUND,           MVT::v8f32, Legal);
     setOperationAction(ISD::FP_EXTEND,          MVT::v8f32, Legal);
 
-    setOperationAction(ISD::TRUNCATE,           MVT::i1, Legal);
+    setOperationAction(ISD::TRUNCATE,           MVT::i1, Custom);
     setOperationAction(ISD::TRUNCATE,           MVT::v16i8, Custom);
     setOperationAction(ISD::TRUNCATE,           MVT::v8i32, Custom);
     setOperationAction(ISD::TRUNCATE,           MVT::v8i1, Custom);
@@ -1374,6 +1374,7 @@ void X86TargetLowering::resetOperationActions() {
     setOperationAction(ISD::CONCAT_VECTORS,     MVT::v16f32,  Custom);
     setOperationAction(ISD::CONCAT_VECTORS,     MVT::v16i32,  Custom);
     setOperationAction(ISD::CONCAT_VECTORS,     MVT::v8i1,    Custom);
+    setOperationAction(ISD::CONCAT_VECTORS,     MVT::v16i1, Legal);
 
     setOperationAction(ISD::SETCC,              MVT::v16i1, Custom);
     setOperationAction(ISD::SETCC,              MVT::v8i1, Custom);
@@ -5425,7 +5426,8 @@ LowerAsSplatVectorLoad(SDValue SrcOp, MVT VT, SDLoc dl, SelectionDAG &DAG) {
 /// rather than undef via VZEXT_LOAD, but we do not detect that case today.
 /// There's even a handy isZeroNode for that purpose.
 static SDValue EltsFromConsecutiveLoads(EVT VT, SmallVectorImpl<SDValue> &Elts,
-                                        SDLoc &DL, SelectionDAG &DAG) {
+                                        SDLoc &DL, SelectionDAG &DAG,
+                                        bool isAfterLegalize) {
   EVT EltVT = VT.getVectorElementType();
   unsigned NumElems = Elts.size();
 
@@ -5461,7 +5463,13 @@ static SDValue EltsFromConsecutiveLoads(EVT VT, SmallVectorImpl<SDValue> &Elts,
   // load of the entire vector width starting at the base pointer.  If we found
   // consecutive loads for the low half, generate a vzext_load node.
   if (LastLoadedElt == NumElems - 1) {
+
+    if (isAfterLegalize &&
+        !DAG.getTargetLoweringInfo().isOperationLegal(ISD::LOAD, VT))
+      return SDValue();
+
     SDValue NewLd = SDValue();
+
     if (DAG.InferPtrAlignment(LDBase->getBasePtr()) >= 16)
       NewLd = DAG.getLoad(VT, DL, LDBase->getChain(), LDBase->getBasePtr(),
                           LDBase->getPointerInfo(),
@@ -6105,7 +6113,7 @@ X86TargetLowering::LowerBUILD_VECTOR(SDValue Op, SelectionDAG &DAG) const {
       V[i] = Op.getOperand(i);
 
     // Check for elements which are consecutive loads.
-    SDValue LD = EltsFromConsecutiveLoads(VT, V, dl, DAG);
+    SDValue LD = EltsFromConsecutiveLoads(VT, V, dl, DAG, false);
     if (LD.getNode())
       return LD;
 
@@ -9066,6 +9074,17 @@ SDValue X86TargetLowering::LowerTRUNCATE(SDValue Op, SelectionDAG &DAG) const {
   MVT VT = Op.getSimpleValueType();
   SDValue In = Op.getOperand(0);
   MVT InVT = In.getSimpleValueType();
+
+  if (VT == MVT::i1) {
+    assert((InVT.isInteger() && (InVT.getSizeInBits() <= 64)) &&
+           "Invalid scalar TRUNCATE operation");
+    In = DAG.getNode(ISD::AND, DL, InVT, In, DAG.getConstant(1, InVT));
+    if (InVT.getSizeInBits() == 64)
+      In = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::i32, In);
+    else if (InVT.getSizeInBits() < 32)
+      In = DAG.getNode(ISD::ANY_EXTEND, DL, MVT::i32, In);
+    return DAG.getNode(X86ISD::TRUNC, DL, VT, In);
+  }
   assert(VT.getVectorNumElements() == InVT.getVectorNumElements() &&
          "Invalid TRUNCATE operation");
 
@@ -15321,7 +15340,7 @@ X86TargetLowering::EmitVAStartSaveXMMRegsWithCustomInserter(
 
   unsigned MOVOpc = Subtarget->hasFp256() ? X86::VMOVAPSmr : X86::MOVAPSmr;
   // In the XMM save block, save all the XMM argument registers.
-  for (int i = 3, e = MI->getNumOperands(); i != e; ++i) {
+  for (int i = 3, e = MI->getNumOperands() - 1; i != e; ++i) {
     int64_t Offset = (i - 3) * 16 + VarArgsFPOffset;
     MachineMemOperand *MMO =
       F->getMachineMemOperand(
@@ -16367,7 +16386,7 @@ static SDValue PerformShuffleCombine(SDNode *N, SelectionDAG &DAG,
   for (unsigned i = 0, e = VT.getVectorNumElements(); i != e; ++i)
     Elts.push_back(getShuffleScalarElt(N, i, DAG, 0));
 
-  return EltsFromConsecutiveLoads(VT, Elts, dl, DAG);
+  return EltsFromConsecutiveLoads(VT, Elts, dl, DAG, true);
 }
 
 /// PerformTruncateCombine - Converts truncate operation to