Fix a bug introduced by rL334850

Summary: All *_sqrt_round_s[s|d] intrinsics should execute a square root on
zeroth element from B (Ops[1]) and insert in to A (Ops[0]), not the other way around.

Reviewers: itaraban, craig.topper

Reviewed By: craig.topper

Subscribers: craig.topper, cfe-commits

Differential Revision: https://reviews.llvm.org/D48288

llvm-svn: 334964

clang/lib/CodeGen/CGBuiltin.cpp

@@ -9907,15 +9907,15 @@ Value *CodeGenFunction::EmitX86BuiltinExpr(unsigned BuiltinID,
                           Intrinsic::x86_avx512_mask_sqrt_ss;
       return Builder.CreateCall(CGM.getIntrinsic(IID), Ops);
     }
-    Value *A = Builder.CreateExtractElement(Ops[0], (uint64_t)0);
+    Value *A = Builder.CreateExtractElement(Ops[1], (uint64_t)0);
     Function *F = CGM.getIntrinsic(Intrinsic::sqrt, A->getType());
     Value *Src = Builder.CreateExtractElement(Ops[2], (uint64_t)0);
     int MaskSize = Ops[3]->getType()->getScalarSizeInBits();
     llvm::Type *MaskTy = llvm::VectorType::get(Builder.getInt1Ty(), MaskSize);
     Value *Mask = Builder.CreateBitCast(Ops[3], MaskTy);
     Mask = Builder.CreateExtractElement(Mask, (uint64_t)0);
     A = Builder.CreateSelect(Mask, Builder.CreateCall(F, {A}), Src);
-    return Builder.CreateInsertElement(Ops[1], A, (uint64_t)0);
+    return Builder.CreateInsertElement(Ops[0], A, (uint64_t)0);
   }
   case X86::BI__builtin_ia32_sqrtpd256:
   case X86::BI__builtin_ia32_sqrtpd: