Fix a bug introduced by rL334850

[tkrupa]

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.

[craig.topper]

LGTM

lib/CodeGen/CGBuiltin.cpp

case X86::BI__builtin_ia32_sqrtss_round_mask: {
     // Support only if the rounding mode is 4 (AKA CUR_DIRECTION),
     // otherwise keep the intrinsic.
     if (CC != 4) {
       Intrinsic::ID IID = BuiltinID == X86::BI__builtin_ia32_sqrtsd_round_mask ?
                           Intrinsic::x86_avx512_mask_sqrt_sd :
                           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:
   case X86::BI__builtin_ia32_sqrtps256:
   case X86::BI__builtin_ia32_sqrtps: {
     Function *F = CGM.getIntrinsic(Intrinsic::sqrt, Ops[0]->getType());
     return Builder.CreateCall(F, {Ops[0]});
   }