Modified InterpBuiltin.cpp and ExprConstant.cpp

Author: markbhasawut

clang/lib/AST/ByteCode/InterpBuiltin.cpp

@@ -2579,10 +2579,30 @@ static bool interp__builtin_elementwise_maxmin(InterpState &S, CodePtr OpPC,
   return true;
 }
 
-static bool interp__builtin_ia32_pmadd(InterpState &S, CodePtr OpPC,
-                                      const CallExpr *Call,
-                                      unsigned BuiltinID) {
-  return true; // TODO: Implement the builtin.
+static bool interp__builtin_ia32_pmadd(
+    InterpState &S, CodePtr OpPC, const CallExpr *Call,
+    llvm::function_ref<APInt(const APSInt &, const APSInt &)> Fn) {
+  assert(Call->getArg(0)->getType()->isVectorType() &&
+         Call->getArg(1)->getType()->isVectorType());
+  const Pointer &RHS = S.Stk.pop<Pointer>();
+  const Pointer &LHS = S.Stk.pop<Pointer>();
+  const Pointer &Dst = S.Stk.peek<Pointer>();
+                                    
+  const auto *VT = Call->getArg(0)->getType()->castAs<VectorType>();
+  PrimType ElemT = *S.getContext().classify(VT->getElementType());
+  unsigned NumElems = VT->getNumElements();
+  bool DestUnsigned = Call->getType()->isUnsignedIntegerOrEnumerationType();
+
+  for (unsigned I = 0; I != NumElems; ++I) {
+    INT_TYPE_SWITCH_NO_BOOL(ElemT, {
+      APSInt Elem1 = LHS.elem<T>(I).toAPSInt();
+      APSInt Elem2 = RHS.elem<T>(I).toAPSInt();
+      Dst.elem<T>(I) = static_cast<T>(APSInt(Fn(Elem1, Elem2), DestUnsigned));
+    });
+  }
+
+  Dst.initializeAllElements();
+  return true;
 }
 
 static bool interp__builtin_ia32_pmul(InterpState &S, CodePtr OpPC,
@@ -3457,12 +3477,20 @@ bool InterpretBuiltin(InterpState &S, CodePtr OpPC, const CallExpr *Call,
   case clang::X86::BI__builtin_ia32_pmaddubsw128:
   case clang::X86::BI__builtin_ia32_pmaddubsw256:
   case clang::X86::BI__builtin_ia32_pmaddubsw512:
-    return true; // TODO: Use interp__builtin_i32_pmadd.
+    return interp__builtin_ia32_pmadd(S, OpPC, Call, 
+  [](const APSInt &LoLHS, const APSInt &HiLHS, const APSInt &LoRHS, const APSInt &HiRHS) {
+    unsigned BitWidth = 2 * LHS.getBitWidth();
+    return (LoLHS.zext(BitWidth) * LoRHS.sext(BitWidth)).sadd_sat((HiLHS.zext(BitWidth) * HiRHS.sext(BitWidth)));
+  });
 
   case clang::X86::BI__builtin_ia32_pmaddwd128:
   case clang::X86::BI__builtin_ia32_pmaddwd256:
   case clang::X86::BI__builtin_ia32_pmaddwd512:
-    return true; // TODO: Use interp__builtin_i32_pmadd.
+    return interp__builtin_ia32_pmadd(S, OpPC, Call, 
+  [](const APSInt &LoLHS, const APSInt &HiLHS, const APSInt &LoRHS, const APSInt &HiRHS) {
+    unsigned BitWidth = 2 * LHS.getBitWidth();
+    return (LoLHS.sext(BitWidth) * LoRHS.sext(BitWidth)) + (HiLHS.sext(BitWidth) * HiRHS.sext(BitWidth));
+  });
 
   case clang::X86::BI__builtin_ia32_pmulhuw128:
   case clang::X86::BI__builtin_ia32_pmulhuw256:

clang/lib/AST/ExprConstant.cpp

@@ -11779,18 +11779,46 @@ bool VectorExprEvaluator::VisitCallExpr(const CallExpr *E) {
     return EvaluateBinOpExpr(llvm::APIntOps::avgCeilU);
 
   case clang::X86::BI__builtin_ia32_pmaddubsw128:
-  case clang::X86::BI__builtin_ia32_pmaddwd128:
   case clang::X86::BI__builtin_ia32_pmaddubsw256:
-  case clang::X86::BI__builtin_ia32_pmaddwd256:
   case clang::X86::BI__builtin_ia32_pmaddubsw512:
-  case clang::X86::BI__builtin_ia32_pmaddwd512:
-    return true; // TODO: Handle __builtin_ia32_pmaddub
-
   case clang::X86::BI__builtin_ia32_pmaddwd128:
   case clang::X86::BI__builtin_ia32_pmaddwd256:
-  case clang::X86::BI__builtin_ia32_pmaddwd512:
-    return true; // TODO: Handle __builtin_ia32_pmadd
-    });
+  case clang::X86::BI__builtin_ia32_pmaddwd512: {
+    APValue SourceLHS, SourceRHS;
+    if (!EvaluateAsRValue(Info, E->getArg(0), SourceLHS) ||
+        !EvaluateAsRValue(Info, E->getArg(1), SourceRHS))
+      return false;
+
+    unsigned SourceLen = SourceLHS.getVectorLength();
+    bool DestUnsigned = DestEltTy->isUnsignedIntegerOrEnumerationType();
+    SmallVector<APValue, 4> ResultElements;
+    ResultElements.reserve(SourceLen / 2);
+
+    for (unsigned EltNum = 0; EltNum < SourceLen; ++EltNum) {
+      const APSInt &LoLHS = SourceLHS.getVectorElt(EltNum).getInt();
+      const APSInt &LoRHS = SourceRHS.getVectorElt(EltNum).getInt();
+      const APSInt &HiLHS = SourceLHS.getVectorElt(EltNum).getInt();
+      const APSInt &HiRHS = SourceRHS.getVectorElt(EltNum).getInt();
+      unsigned BitWidth = 2 * LHS.getBitWidth();
+
+      switch (E->getBuiltinCallee()) {
+      case clang::X86::BI__builtin_ia32_pmaddubsw128:
+      case clang::X86::BI__builtin_ia32_pmaddubsw256:
+      case clang::X86::BI__builtin_ia32_pmaddubsw512:
+        ResultElements.push_back(
+          APValue(APSInt((LoLHS.zext(BitWidth) * LoRHS.sext(BitWidth)).sadd_sat((HiLHS.zext(BitWidth) * HiRHS.sext(BitWidth)))), DestUnsigned));
+        break;
+      case clang::X86::BI__builtin_ia32_pmaddwd128:
+      case clang::X86::BI__builtin_ia32_pmaddwd256:
+      case clang::X86::BI__builtin_ia32_pmaddwd512:
+        ResultElements.push_back(
+            APValue(APSInt((LoLHS.sext(BitWidth) * LoRHS.sext(BitWidth)) + (HiLHS.sext(BitWidth) * HiRHS.sext(BitWidth))), DestUnsigned));
+        break;
+      }
+    }
+
+    return Success(APValue(ResultElements.data(), ResultElements.size()), E);
+  }
 
   case clang::X86::BI__builtin_ia32_pmulhuw128:
   case clang::X86::BI__builtin_ia32_pmulhuw256: