[turbofan] Added the RoundInt64ToFloat64 instruction to TurboFan.

src/compiler/machine-operator.h

 // Copyright 2013 the V8 project authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #ifndef V8_COMPILER_MACHINE_OPERATOR_H_
 #define V8_COMPILER_MACHINE_OPERATOR_H_
 
 #include "src/base/flags.h"
 #include "src/compiler/machine-type.h"
 
@@ skipping 174 matching lines @@
   // in the target type and are *not* defined for other inputs.
   // Use narrowing change operators only when there is a static guarantee that
   // the input value is representable in the target value.
   const Operator* ChangeFloat32ToFloat64();
   const Operator* ChangeFloat64ToInt32();   // narrowing
   const Operator* ChangeFloat64ToUint32();  // narrowing
   const Operator* ChangeInt32ToFloat64();
   const Operator* ChangeInt32ToInt64();
   const Operator* ChangeUint32ToFloat64();
   const Operator* ChangeUint32ToUint64();
+  const Operator* ChangeInt64ToFloat64();

[titzer, 2015/10/30 09:44:06]

I think we should call this "RoundInt64ToFloat64" since there is a possible loss
of precision.

[ahaas, 2015/11/02 10:35:37]

Done.

 
   // These operators truncate numbers, both changing the representation of
   // the number and mapping multiple input values onto the same output value.
   const Operator* TruncateFloat64ToFloat32();
   const Operator* TruncateFloat64ToInt32(TruncationMode);
   const Operator* TruncateInt64ToInt32();
 
   // These operators reinterpret the bits of a floating point number as an
   // integer and vice versa.
   const Operator* BitcastFloat32ToInt32();
@@ skipping 111 matching lines @@
 };
 
 
 DEFINE_OPERATORS_FOR_FLAGS(MachineOperatorBuilder::Flags)
 
 }  // namespace compiler
 }  // namespace internal
 }  // namespace v8
 
 #endif  // V8_COMPILER_MACHINE_OPERATOR_H_