https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107591
--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> --- Perhaps before we try to map MULT_EXPR into some irange/frange op the usual way, while we still have access to gimple statement check if it is MULT_EXPR with the same operands and use a different artificial unary op for pow2. We could do this for PLUS_EXPR x + x as well and handle it as 2 * x if we don't do that already or if match.pd doesn't already canonicalize x + x that way.
