RISC-V / GCC 10: invalid operation not set in acos and asin

[Joseph Myers]

On Wed, 13 May 2020, Adhemerval Zanella via Libc-alpha wrote:

> Is is an invalid optimization? Because:
> 
> double foo (double x)
> {
>   x = NAN;
>   x = x / x;
>   return x;
> }

Optimizing away NAN / NAN is valid, because that doesn't raise any 
exceptions.  Optimizing away 0.0 / 0.0 or INFINITY / INFINITY isn't, 
because that should raise "invalid".

-- 
Joseph S. Myers
joseph@codesourcery.com