[libc-commits] [libc] [libc][math] Fix missing underflow exception in DyadicFloat::generic_as (PR #186734)

[via libc-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-libc

Author: Anonmiraj (AnonMiraj)

<details>
<summary>Changes</summary>

The `generic_as` function in `dyadic_float.h` had a missing `underflow = true` at the exact boundary where `unbiased_exp == -EXP_BIAS - FRACTION_LEN`.
At this boundary, the mantissa MSB maps exactly to the round bit, so out_mantissa is 0 and the result can only be 0 or min_subnormal. The value is at most min_subnormal / 2, so it is always tiny and always inexact  `underflow` must be signaled. The < case and the general subnormal range both already set underflow = true this boundary case was the only gap.

this specifically fix this error in the erfcf16 function 
```
Extracted erfcf16.cpp.o from archive for linking
Running exhaustive check in --rndn mode...
Missing underflow exception for x=0x1.eacp+1 (y=0x1p-24)
```

this fix may also apply to other bfloat16 missing exceptions (@<!-- -->Sukumarsawant)

part of: #<!-- -->186483 

CC: @<!-- -->lntue 

---
Full diff: https://github.com/llvm/llvm-project/pull/186734.diff


1 Files Affected:

- (modified) libc/src/__support/FPUtil/dyadic_float.h (+1) 


``````````diff
diff --git a/libc/src/__support/FPUtil/dyadic_float.h b/libc/src/__support/FPUtil/dyadic_float.h
index cc0710fbf7b02..018429b30272a 100644
--- a/libc/src/__support/FPUtil/dyadic_float.h
+++ b/libc/src/__support/FPUtil/dyadic_float.h
@@ -219,6 +219,7 @@ template <size_t Bits> struct DyadicFloat {
       underflow = true;
     } else if (unbiased_exp == -FPBits::EXP_BIAS - FPBits::FRACTION_LEN) {
       round = true;
+      underflow = true;
       MantissaType sticky_mask = (MantissaType(1) << (Bits - 1)) - 1;
       sticky = (mantissa & sticky_mask) != 0;
     } else {

``````````

</details>


https://github.com/llvm/llvm-project/pull/186734