Use signaling NaN for holes in fixed double arrays.

src/objects-inl.h

Index: src/objects-inl.h
diff --git a/src/objects-inl.h b/src/objects-inl.h
index 43db240fd9c3d71538411c63fd468d741aaea534..2bfdedd87f4398938b5e71f9ce8faaba69fd1737 100644
--- a/src/objects-inl.h
+++ b/src/objects-inl.h
@@ -2289,19 +2289,12 @@ void FixedArray::set(int index, Object* value) {
 
 
 inline bool FixedDoubleArray::is_the_hole_nan(double value) {
-  return bit_cast<uint64_t, double>(value) == kHoleNanInt64;
+  return bit_cast<uint64_t>(value) == kHoleNanInt64;
 }
 
 
 inline double FixedDoubleArray::hole_nan_as_double() {
-  return bit_cast<double, uint64_t>(kHoleNanInt64);
-}
-
-
-inline double FixedDoubleArray::canonical_not_the_hole_nan_as_double() {
-  DCHECK(bit_cast<uint64_t>(base::OS::nan_value()) != kHoleNanInt64);
-  DCHECK((bit_cast<uint64_t>(base::OS::nan_value()) >> 32) != kHoleNanUpper32);
-  return base::OS::nan_value();
+  return bit_cast<double>(kHoleNanInt64);
 }
 
 
@@ -2336,7 +2329,9 @@ void FixedDoubleArray::set(int index, double value) {
   DCHECK(map() != GetHeap()->fixed_cow_array_map() &&
          map() != GetHeap()->fixed_array_map());
   int offset = kHeaderSize + index * kDoubleSize;
-  if (std::isnan(value)) value = canonical_not_the_hole_nan_as_double();
+  if (std::isnan(value)) {
+    value = std::numeric_limits<double>::quiet_NaN();
+  }
   WRITE_DOUBLE_FIELD(this, offset, value);
 }