https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109933
--- Comment #13 from Patrick O'Neill <patrick at rivosinc dot com> --- Created attachment 56487 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=56487&action=edit Proposed fix Hi Rory, Sorry that this slipped off my plate for way too long. I just got around to refactoring the surrounding code on trunk and revised your patch to fix this for both test_and_set along with inline subword amo sequences. When you have the time can you please test this using your big-endian setup? I'll handle the little-endian testing. Patrick
