https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69857
--- Comment #4 from David Binderman <dcb314 at hotmail dot com> --- (In reply to ktkachov from comment #3) > Agree that gcc should warn here. > As for the suspicious return itself, from what I can see it has the effect > of overly restricting generation of the LDRD/STRD instructions on arm. > Removing it is probably the right thing to do (and passes a first round of > testing for me). > > However, since the code it was blocking has not been tested so far (because > it was never executed) we don't know what bugs (if any) are lurking there, > so removing it should wait for GCC 7 at this point What would be enough testing to ensure the change would get into gcc 6 ? Maybe a gcc bootstrap build or a linux kernel build. Maybe a full build of all the packages in Fedora / OpenSuse / Debian / Ubuntu. I don't have ARM here, but this might be enough excuse to get access to the compile farm or to some other ARM server.
