diff --git a/src/coreclr/jit/codegenwasm.cpp b/src/coreclr/jit/codegenwasm.cpp
index 0bf6b32b214fa5..4821043729bac8 100644
--- a/src/coreclr/jit/codegenwasm.cpp
+++ b/src/coreclr/jit/codegenwasm.cpp
@@ -1883,12 +1883,22 @@ void CodeGen::genCodeForBitCast(GenTreeOp* tree)
         case PackTypes(TYP_DOUBLE, TYP_LONG):
             ins = INS_f64_reinterpret_i64;
             break;
+
+        // Same-size bitcasts are no-ops on the wasm value stack. PackTypes normalizes
+        // TYP_REF/TYP_BYREF to TYP_I_IMPL, so this covers all INT/REF/BYREF combos on wasm32
+        // and LONG/REF/BYREF on wasm64.
+        case PackTypes(TYP_INT, TYP_INT):
+        case PackTypes(TYP_LONG, TYP_LONG):
+            break;
+
         default:
             unreached();
-            break;
     }
 
-    GetEmitter()->emitIns(ins);
+    if (ins != INS_none)
+    {
+        GetEmitter()->emitIns(ins);
+    }
     WasmProduceReg(tree);
 }
 
diff --git a/src/coreclr/jit/valuenum.cpp b/src/coreclr/jit/valuenum.cpp
index e657b9dde198f5..6e5460df0766a5 100644
--- a/src/coreclr/jit/valuenum.cpp
+++ b/src/coreclr/jit/valuenum.cpp
@@ -38,19 +38,15 @@ struct FloatTraits
     // Notes:
     //    "Default" NaN value returned by expression 0.0f / 0.0f on x86/x64 has
     //    different binary representation (0xffc00000) than NaN on
-    //    ARM32/ARM64/LoongArch64 (0x7fc00000).
+    //    ARM32/ARM64/LoongArch64/RISC-V 64/WASM (0x7fc00000).
 
     static float NaN()
     {
 #if defined(TARGET_XARCH)
         unsigned bits = 0xFFC00000u;
-#elif defined(TARGET_ARMARCH) || defined(TARGET_LOONGARCH64) || defined(TARGET_RISCV64)
+#elif defined(TARGET_ARMARCH) || defined(TARGET_LOONGARCH64) || defined(TARGET_RISCV64) || defined(TARGET_WASM)
+        // WASM uses the canonical NaN payload, matching ARM.
         unsigned bits = 0x7FC00000u;
-#elif defined(TARGET_WASM)
-        // TODO-WASM: this may prove tricker than it seems since there are two possible "canonical"
-        // NaN values. We may need to introduce a new "unknown" value to be returned here.
-        NYI_WASM("FloatTraits::NaN");
-        unsigned bits = 0;
 #else
 #error Unsupported or unset target architecture
 #endif
@@ -69,19 +65,15 @@ struct DoubleTraits
     // Notes:
     //    "Default" NaN value returned by expression 0.0 / 0.0 on x86/x64 has
     //    different binary representation (0xfff8000000000000) than NaN on
-    //    ARM32/ARM64/LoongArch64 (0x7ff8000000000000).
+    //    ARM32/ARM64/LoongArch64/RISC-V 64/WASM (0x7ff8000000000000).
 
     static double NaN()
     {
 #if defined(TARGET_XARCH)
         unsigned long long bits = 0xFFF8000000000000ull;
-#elif defined(TARGET_ARMARCH) || defined(TARGET_LOONGARCH64) || defined(TARGET_RISCV64)
+#elif defined(TARGET_ARMARCH) || defined(TARGET_LOONGARCH64) || defined(TARGET_RISCV64) || defined(TARGET_WASM)
+        // WASM uses the canonical NaN payload, matching ARM.
         unsigned long long bits = 0x7FF8000000000000ull;
-#elif defined(TARGET_WASM)
-        // TODO-WASM: this may prove tricker than it seems since there are two possible "canonical"
-        // NaN values. We may need to introduce a new "unknown" value to be returned here.
-        NYI_WASM("DoubleTraits::NaN");
-        unsigned long long bits = 0;
 #else
 #error Unsupported or unset target architecture
 #endif
@@ -96,7 +88,7 @@ struct DoubleTraits
 // FpAdd: Computes value1 + value2
 //
 // Return Value:
-//    TFpTraits::NaN() - If target ARM32/ARM64 and result value is NaN
+//    TFpTraits::NaN() - If target is ARM32/ARM64/LoongArch64/RISC-V 64/WASM and result value is NaN
 //    value1 + value2  - Otherwise
 //
 // Notes:
@@ -105,7 +97,7 @@ struct DoubleTraits
 template <typename TFp, typename TFpTraits>
 TFp FpAdd(TFp value1, TFp value2)
 {
-#if defined(TARGET_ARMARCH) || defined(TARGET_LOONGARCH64) || defined(TARGET_RISCV64)
+#if defined(TARGET_ARMARCH) || defined(TARGET_LOONGARCH64) || defined(TARGET_RISCV64) || defined(TARGET_WASM)
     // If [value1] is negative infinity and [value2] is positive infinity
     //   the result is NaN.
     // If [value1] is positive infinity and [value2] is negative infinity
@@ -123,7 +115,7 @@ TFp FpAdd(TFp value1, TFp value2)
             return TFpTraits::NaN();
         }
     }
-#endif // TARGET_ARMARCH || TARGET_LOONGARCH64 || TARGET_RISCV64
+#endif // TARGET_ARMARCH || TARGET_LOONGARCH64 || TARGET_RISCV64 || TARGET_WASM
 
     return value1 + value2;
 }
@@ -132,7 +124,7 @@ TFp FpAdd(TFp value1, TFp value2)
 // FpSub: Computes value1 - value2
 //
 // Return Value:
-//    TFpTraits::NaN() - If target ARM32/ARM64 and result value is NaN
+//    TFpTraits::NaN() - If target is ARM32/ARM64/LoongArch64/RISC-V 64/WASM and result value is NaN
 //    value1 - value2  - Otherwise
 //
 // Notes:
@@ -141,7 +133,7 @@ TFp FpAdd(TFp value1, TFp value2)
 template <typename TFp, typename TFpTraits>
 TFp FpSub(TFp value1, TFp value2)
 {
-#if defined(TARGET_ARMARCH) || defined(TARGET_LOONGARCH64) || defined(TARGET_RISCV64)
+#if defined(TARGET_ARMARCH) || defined(TARGET_LOONGARCH64) || defined(TARGET_RISCV64) || defined(TARGET_WASM)
     // If [value1] is positive infinity and [value2] is positive infinity
     //   the result is NaN.
     // If [value1] is negative infinity and [value2] is negative infinity
@@ -159,7 +151,7 @@ TFp FpSub(TFp value1, TFp value2)
             return TFpTraits::NaN();
         }
     }
-#endif // TARGET_ARMARCH || TARGET_LOONGARCH64 || TARGET_RISCV64
+#endif // TARGET_ARMARCH || TARGET_LOONGARCH64 || TARGET_RISCV64 || TARGET_WASM
 
     return value1 - value2;
 }
@@ -168,7 +160,7 @@ TFp FpSub(TFp value1, TFp value2)
 // FpMul: Computes value1 * value2
 //
 // Return Value:
-//    TFpTraits::NaN() - If target ARM32/ARM64 and result value is NaN
+//    TFpTraits::NaN() - If target is ARM32/ARM64/LoongArch64/RISC-V 64/WASM and result value is NaN
 //    value1 * value2  - Otherwise
 //
 // Notes:
@@ -177,7 +169,7 @@ TFp FpSub(TFp value1, TFp value2)
 template <typename TFp, typename TFpTraits>
 TFp FpMul(TFp value1, TFp value2)
 {
-#if defined(TARGET_ARMARCH) || defined(TARGET_LOONGARCH64) || defined(TARGET_RISCV64)
+#if defined(TARGET_ARMARCH) || defined(TARGET_LOONGARCH64) || defined(TARGET_RISCV64) || defined(TARGET_WASM)
     // From the ECMA standard:
     //
     // If [value1] is zero and [value2] is infinity
@@ -193,7 +185,7 @@ TFp FpMul(TFp value1, TFp value2)
     {
         return TFpTraits::NaN();
     }
-#endif // TARGET_ARMARCH || TARGET_LOONGARCH64 || TARGET_RISCV64
+#endif // TARGET_ARMARCH || TARGET_LOONGARCH64 || TARGET_RISCV64 || TARGET_WASM
 
     return value1 * value2;
 }
@@ -202,7 +194,7 @@ TFp FpMul(TFp value1, TFp value2)
 // FpDiv: Computes value1 / value2
 //
 // Return Value:
-//    TFpTraits::NaN() - If target ARM32/ARM64 and result value is NaN
+//    TFpTraits::NaN() - If target is ARM32/ARM64/LoongArch64/RISC-V 64/WASM and result value is NaN
 //    value1 / value2  - Otherwise
 //
 // Notes:
@@ -211,7 +203,7 @@ TFp FpMul(TFp value1, TFp value2)
 template <typename TFp, typename TFpTraits>
 TFp FpDiv(TFp dividend, TFp divisor)
 {
-#if defined(TARGET_ARMARCH) || defined(TARGET_LOONGARCH64) || defined(TARGET_RISCV64)
+#if defined(TARGET_ARMARCH) || defined(TARGET_LOONGARCH64) || defined(TARGET_RISCV64) || defined(TARGET_WASM)
     // From the ECMA standard:
     //
     // If [dividend] is zero and [divisor] is zero
@@ -228,7 +220,7 @@ TFp FpDiv(TFp dividend, TFp divisor)
     {
         return TFpTraits::NaN();
     }
-#endif // TARGET_ARMARCH || TARGET_LOONGARCH64 || TARGET_RISCV64
+#endif // TARGET_ARMARCH || TARGET_LOONGARCH64 || TARGET_RISCV64 || TARGET_WASM
 
     return dividend / divisor;
 }