https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108375
--- Comment #6 from Martin Uecker <muecker at gwdg dot de> --- Created attachment 54458 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54458&action=edit preliminary patch against trunk Here is a preliminary patch. It takes a bit from type_common to cache negative results for records / unions (i.e. not a vm-type). Performance might be ok, although I think an approach where the FE sets the bit for records which are known to be variaby modified and all others are not might be preferable.
