https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80530
wilco at gcc dot gnu.org changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |wilco at gcc dot gnu.org
--- Comment #2 from wilco at gcc dot gnu.org ---
Basically what happens we claim the builtin exists and then when rsqrt is to be
expanded we decide not to emit it. There is no alternative (unlike div and sqrt
which can fall back to their original instruction), so it fails.
aarch64_emit_approx_sqrt needs to use use_rsqrt_p when recp is true rather than
using a different condition.
