https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117011
Jeffrey A. Law <law at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Resolution|--- |WONTFIX
Status|UNCONFIRMED |RESOLVED
--- Comment #2 from Jeffrey A. Law <law at gcc dot gnu.org> ---
Jovan, I don't think I'd worry much about this and defer it until such time as
someone actually wants to expose zicond + short-forward-branch together.
As I noted in the meeting today, we've actually got both in our design, but we
generally try to steer towards zicond and use our SFB to handle things that are
either missed by zicond (usually because they involve type/mode changes) or for
cases where we're running more generic code that may have been compiled without
zicond. Thus we don't even expose that we have SFB to the compiler for the
Ventana designs.
