https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91384
--- Comment #3 from Uroš Bizjak <ubizjak at gmail dot com> --- (In reply to Jakub Jelinek from comment #1) > Started with r223689. Though, generally that change looks like a useful > GIMPLE canonicalization. FYI, -fno-tree-forwprop generates expected code. > Can't this be done in a peephole2? Sure, but this looks like a more generic missed optimization problem, where solution would benefit several targets.
