[libc-commits] [libc] [libc][math] Update getpayload and fmul with NaN inputs. (PR #99812)

[via libc-commits]

================
@@ -50,19 +50,19 @@ mul(InType x, InType y) {
         raise_except_if_required(FE_INVALID);
 
       if (x_bits.is_quiet_nan()) {
-        InStorageType x_payload = static_cast<InStorageType>(getpayload(x));
-        if ((x_payload & ~(OutFPBits::FRACTION_MASK >> 1)) == 0)
-          return OutFPBits::quiet_nan(x_bits.sign(),
-                                      static_cast<OutStorageType>(x_payload))
-              .get_val();
+        InStorageType x_payload = x_bits.get_mantissa();
+        x_payload >>= (InFPBits::FRACTION_LEN - OutFPBits::FRACTION_LEN);
+        return OutFPBits::quiet_nan(x_bits.sign(),
+                                    static_cast<OutStorageType>(x_payload))
+            .get_val();
----------------
lntue wrote:

Most of the nan-propagations, especially binary operations, seem to be best-effort, and a bit underspecified in both IEEE 754 and C standards.  

https://github.com/llvm/llvm-project/pull/99812