diff --git a/Changelog b/Changelog
index b5ca3e8..a3add38 100644
--- a/Changelog
+++ b/Changelog
@@ -1,3 +1,6 @@
+2020-02-16:
+  Add support for quiet and signaling NaNs to ieee header.
+
 2019-10-31:
   Add support for xtensa architecture.
   Add support for nios2 architecture.
diff --git a/double-conversion/ieee.h b/double-conversion/ieee.h
index 8c3b862..3c2a597 100644
--- a/double-conversion/ieee.h
+++ b/double-conversion/ieee.h
@@ -45,6 +45,7 @@
   static const uint64_t kExponentMask = DOUBLE_CONVERSION_UINT64_2PART_C(0x7FF00000, 00000000);
   static const uint64_t kSignificandMask = DOUBLE_CONVERSION_UINT64_2PART_C(0x000FFFFF, FFFFFFFF);
   static const uint64_t kHiddenBit = DOUBLE_CONVERSION_UINT64_2PART_C(0x00100000, 00000000);
+  static const uint64_t kQuietNanBit = DOUBLE_CONVERSION_UINT64_2PART_C(0x00080000, 00000000);
   static const int kPhysicalSignificandSize = 52;  // Excludes the hidden bit.
   static const int kSignificandSize = 53;
   static const int kExponentBias = 0x3FF + kPhysicalSignificandSize;
@@ -148,6 +149,15 @@
         ((d64 & kSignificandMask) != 0);
   }
 
+  bool IsQuietNan() const {
+    return IsNan() && ((AsUint64() & kQuietNanBit) != 0);
+  }
+
+  bool IsSignalingNan() const {
+    return IsNan() && ((AsUint64() & kQuietNanBit) == 0);
+  }
+
+
   bool IsInfinite() const {
     uint64_t d64 = AsUint64();
     return ((d64 & kExponentMask) == kExponentMask) &&
@@ -266,6 +276,7 @@
   static const uint32_t kExponentMask = 0x7F800000;
   static const uint32_t kSignificandMask = 0x007FFFFF;
   static const uint32_t kHiddenBit = 0x00800000;
+  static const uint32_t kQuietNanBit = 0x00400000;
   static const int kPhysicalSignificandSize = 23;  // Excludes the hidden bit.
   static const int kSignificandSize = 24;
 
@@ -324,6 +335,15 @@
         ((d32 & kSignificandMask) != 0);
   }
 
+  bool IsQuietNan() const {
+    return IsNan() && ((AsUint32() & kQuietNanBit) != 0);
+  }
+
+  bool IsSignalingNan() const {
+    return IsNan() && ((AsUint32() & kQuietNanBit) == 0);
+  }
+
+
   bool IsInfinite() const {
     uint32_t d32 = AsUint32();
     return ((d32 & kExponentMask) == kExponentMask) &&
diff --git a/test/cctest/test-ieee.cc b/test/cctest/test-ieee.cc
index aebecf5..312eccd 100644
--- a/test/cctest/test-ieee.cc
+++ b/test/cctest/test-ieee.cc
@@ -26,6 +26,7 @@
 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 #include <stdlib.h>
+#include <limits>
 
 #include "cctest.h"
 #include "double-conversion/diy-fp.h"
@@ -444,3 +445,19 @@
   CHECK_EQ(-Double::Infinity(),
            Double(DOUBLE_CONVERSION_UINT64_2PART_C(0xffefffff, ffffffff)).PreviousDouble());
 }
+
+TEST(SignalingNan) {
+  Double nan(Double::NaN());
+  CHECK(nan.IsNan());
+  CHECK(nan.IsQuietNan());
+  CHECK(Double(std::numeric_limits<double>::quiet_NaN()).IsQuietNan());
+  CHECK(Double(std::numeric_limits<double>::signaling_NaN()).IsSignalingNan());
+}
+
+TEST(SignalingNanSingle) {
+  Single nan(Single::NaN());
+  CHECK(nan.IsNan());
+  CHECK(nan.IsQuietNan());
+  CHECK(Single(std::numeric_limits<float>::quiet_NaN()).IsQuietNan());
+  CHECK(Single(std::numeric_limits<float>::signaling_NaN()).IsSignalingNan());
+}