https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99863
--- Comment #16 from Jakub Jelinek <jakub at gcc dot gnu.org> --- (In reply to rguent...@suse.de from comment #15) > On Thu, 1 Apr 2021, jakub at gcc dot gnu.org wrote: > > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99863 > > > > --- Comment #14 from Jakub Jelinek <jakub at gcc dot gnu.org> --- > > Though, we don't add the multiple stores case to active_local_stores and all > > the replace_read calls are from the active_local_stores chain. > > So I think the above patch should DTRT. > > I suppose it already fails elsewhere if fixed_regs_live is not > available? No. But the way I wrote the above patch, if fixed_regs_live is not available, it will punt if any hard regs are live in the considered sequence.
