https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122662

--- Comment #6 from Wilco <wilco at gcc dot gnu.org> ---
(In reply to Mohamed from comment #4)

> One thing that was noticed also in gcc-12 generated assembly(above) built on
> arm-v8, that the binary generated is triggering FE_INVALID, but due to
> something else.

Floating point compares always trigger FE_INVALID on signaling NaNs. It's
exactly the same issue as x86. GCC does not correctly handle signaling NaNs for
any of the isinf/isnan/isnormal/isfinite/iszero/isdenormal/fpclassify
functions. The first 4 are fixed on AArch64 trunk (and a few other targets also
fixed isinf or isnan IIRC).

Reply via email to