https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102865
--- Comment #3 from Heather A <alpha+gcc at alphaservcomputing dot solutions> --- Ha, okay. I figured it out. It was totally bad on my side. https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=7cbc870c495cebc61f5d0ebb975856c207a42fab resolved this issue. I was going insane trying to figure out why ix86_scalar_mode_supported_p did not have a case like you just pointed out, but it was because this was actually off of gcc 11.2 with some backported functionality (from Iain Sandoe's Darwin aarch64 work). I thought it reproduced clearly on master, but I guess I messed up that test. Well, that's embarrassing. Thanks! And sorry about that!
