[turbofan] Properly lower NumberTan to Float64Tan.

src/compiler/simplified-lowering.cc

 // Copyright 2014 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.
 
 #include "src/compiler/simplified-lowering.h"
 
 #include <limits>
 
 #include "src/address-map.h"
 #include "src/base/bits.h"
@@ skipping 1521 matching lines @@
       case IrOpcode::kNumberAtan:
       case IrOpcode::kNumberAtanh:
       case IrOpcode::kNumberCos:
       case IrOpcode::kNumberExp:
       case IrOpcode::kNumberExpm1:
       case IrOpcode::kNumberLog:
       case IrOpcode::kNumberLog1p:
       case IrOpcode::kNumberLog2:
       case IrOpcode::kNumberLog10:
       case IrOpcode::kNumberCbrt:
-      case IrOpcode::kNumberSin: {
+      case IrOpcode::kNumberSin:
+      case IrOpcode::kNumberTan: {
         VisitUnop(node, UseInfo::TruncatingFloat64(),
                   MachineRepresentation::kFloat64);
         if (lower()) NodeProperties::ChangeOp(node, Float64Op(node));
         return;
       }
       case IrOpcode::kNumberRound: {
         VisitUnop(node, UseInfo::TruncatingFloat64(),
                   MachineRepresentation::kFloat64);
         if (lower()) DeferReplacement(node, lowering->Float64Round(node));
         return;
@@ skipping 1501 matching lines @@
         isolate(), graph()->zone(), callable.descriptor(), 0, flags,
         Operator::kNoProperties);
     to_number_operator_.set(common()->Call(desc));
   }
   return to_number_operator_.get();
 }
 
 }  // namespace compiler
 }  // namespace internal
 }  // namespace v8