https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89150
--- Comment #5 from rguenther at suse dot de <rguenther at suse dot de> --- On February 1, 2019 7:31:35 PM GMT+01:00, "jakub at gcc dot gnu.org" <gcc-bugzi...@gcc.gnu.org> wrote: >https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89150 > >Jakub Jelinek <jakub at gcc dot gnu.org> changed: > > What |Removed |Added >---------------------------------------------------------------------------- > CC| |jakub at gcc dot gnu.org > >--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> --- >Are bitmap_elements when not in tree form ever walked from anything but >the >first bitmap_element? bitmap_ggc_free is deletable, so shouldn't be >marked at >all. As current is marked skip and first is not, I think it should >have no >advantage to have chain_prev ("%h.prev"). Well, chain_prev would be better than not gty((skip)) which for list form we could have anyways. That said, a custom walker for bitmap_head might be optimal.
