https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96375
--- Comment #4 from Christophe Lyon <clyon at gcc dot gnu.org> --- (In reply to Andrea Corallo from comment #3) > "clyon at gcc dot gnu.org" <gcc-bugzi...@gcc.gnu.org> writes: > > Hi, > > Hi, > > > It does fix the FAIL, thanks. > > Thanks for testing it! > > > I suspect you also want to add -mthumb to > > check_effective_target_arm_v8_1_lob_ok for consistency. > > isn't the patch already doing this? Hmmm right, I probably missed that part when testing manually, sorry for the noise. > > > In practice, how do you exercise these tests given that with the arm-eabi > > configurations I'm testing they are unsupported because I override -mcpu? > > Not sure I understand, what is the problem if the test is not supported > with a specific -mcpu? > Just that I have no config where lob[16].c are supported because when running arm-none-eabi tests I override -mcpu. When I don't override -mcpu, I don't enable multilibs which in turn causes some effective-target checks failures. Hence I'm wondering what toolchain settings you are using to run these new tests?