This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: calculate_global_regs_live
> In gcc-3.1, update_life_info now calls calculate_global_regs_live in a
> loop, so it is called multiple times with one update_life_info call.
> The calculate_global_regs_live comment you quoted appears, but it seems
> to be the same algorithm, so it appears to be just a documentation
> improvement.
Yes, it was added by a comment-only patch by Mark Mitchell.
> On mainline, there is code in update_life_info to clear the fields in
> question at the end of the loop before calling
> calculate_globals_regs_live again. So it appears that someone has
> already noticed a similar problem to yours.
Oh! I managed to miss it on mainline (the problem I see is on the 3.3
branch). The code was added by
2003-01-30 Richard Henderson <rth@redhat.com>
* flow.c (update_life_info): Zap life info after cleanup_cfg.
(regno_uninitialized): Use correct live at function entry set.
(regno_clobbered_at_setjmp): Likewise.
but, according to the comment, not for the problem I ran into. I'll update
the comment on mainline.
Do you think it is safe to backport the patch (or only part of it) to the 3.3
branch? It's PR optimization/12158, originating from Debian.
Thanks.
--
Eric Botcazou