This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patches] better tested jump killer patch
- To: Jan Hubicka <jh at suse dot cz>
- Subject: Re: [patches] better tested jump killer patch
- From: Jan Hubicka <hubicka at atrey dot karlin dot mff dot cuni dot cz>
- Date: Fri, 20 Jul 2001 22:16:50 +0200
- Cc: gcc-patches at gcc dot gnu dot org, rth at cygnus dot com, patches at x86-64 dot org
- References: <20010720174513.F22463@atrey.karlin.mff.cuni.cz>
> Hi,
> I hope I've now solved the problems brought by killing jump_optimize_minimal,
> (it was much harder that I've expected) so I would love to proceed by killing
> jump_optimize entirely. This is the updated patch I've sent some time ago for
> current tree.
>
> I've gave it additional checking (bootstrap+regtest) on mips/ppc/i686 and found
> no problems except the set_noop_p issue I've sent patch today. On PPC it fixes
> the flow1 failure by not leaving the ADDR_VEC in wrong place of instruction
> stream. The Sparc bootstrap+regtest is in progress. It has finished stage2
> now, so I am optimistic.
>
> OK to install now? (assuming that I find no additional problems on sparc?)
Note that the sparc testing has passed too.
Honza
>
> Honza
>
> Ne čec 15 22:45:37 CEST 2001 Jan Hubicka <jh@suse.cz>
>
> * jump.c: Update comments.
> (delete_barrier_successors, delete_unreferenced_labels,
> delete_noop_moves, tension_vector_labels, delete_from_jump_chain,
> delete_labelref_insn, redirect_tablejump, jump_optimize_1,
> jump_optimize, jump_optimize_minimal): Kill.
> (rebuild_jump_labels): Move code from jump_optimize_1.
> (purge_line_number_notes): Likewise.
> (copy_loop_headers): Likewise.
> * reg-stack.c: Update comment.
> * stmt.c: Likewise.
> * rtl.h (jump_optimize, jump_optimize_1, JUMP_NOOP_MOVES,
> JUMP_AFTER_REGSCAN): Kill.
> * toplev.c (rest_of_compilation): Use cleanup_cfg, call
> purge_line_number_notes and copy_loop_headers.