https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124362
--- Comment #11 from Jeffrey A. Law <law at gcc dot gnu.org> --- I've reviewed the .optimized diffs as well. Ultimately everything looks good. There was brief moment when I thought I'd found something, but it was just canonicalization changing the ordering of operands due to changes in SSA_NAME_VERSIONs. So thinking this is backend/target at this point. Given the DOM change was fixing a bogus warning, I'm inclined to revert it and revisit once stage1 opens again.
