https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65215
--- Comment #7 from Thomas Preud'homme <thopre01 at gcc dot gnu.org> --- (In reply to Jakub Jelinek from comment #6) > I can certainly remove that hunk from the patch, if the expander and other > passes handle it well. The test can stay I guess. Things are at least working on x86 (obviously), ARM and SPARC (after PR61320's fix). Also this code is there since a long time without any bug report problems due to unalignment. One question about the patch: is there a reason not to use n->range instead of GET_MODE_BITSIZE (TYPE_MODE (load_type))? Thanks for finding and fixing this. Best regards.