This is the mail archive of the
mailing list for the GCC project.
Re: [4/9] Remove AND_HARD_REG_SET
- From: Jeff Law <law at redhat dot com>
- To: gcc-patches at gcc dot gnu dot org, richard dot sandiford at arm dot com
- Date: Mon, 9 Sep 2019 11:38:39 -0600
- Subject: Re: [4/9] Remove AND_HARD_REG_SET
- References: <email@example.com> <firstname.lastname@example.org>
On 9/9/19 9:59 AM, Richard Sandiford wrote:
> Use "x &= y" instead of "AND_HARD_REG_SET (x, y)" (or just "x & y"
> if the result is a temporary).
> 2019-09-09 Richard Sandiford <email@example.com>
> * hard-reg-set.h (HARD_REG_SET::operator&): New function.
> (HARD_REG_SET::operator&): Likewise.
> (AND_HARD_REG_SET): Delete.
> * caller-save.c (setup_save_areas): Use "&" instead of
> (save_call_clobbered_regs): Likewise.
> * config/gcn/gcn.c (gcn_md_reorg): Likewise.
> * config/m32c/m32c.c (reduce_class): Likewise.
> * config/rs6000/rs6000.c (rs6000_register_move_cost): Likewise.
> * final.c (get_call_reg_set_usage): Likewise.
> * ira-color.c (add_allocno_hard_regs_to_forest): Likewise.
> (setup_left_conflict_sizes_p): Likewise.
> * ira-conflicts.c (print_allocno_conflicts): Likewise.
> (ira_build_conflicts): Likewise.
> * ira-costs.c (restrict_cost_classes): Likewise.
> * ira.c (setup_stack_reg_pressure_class, setup_class_translate_array)
> (setup_reg_class_relations): Likewise.
> * reginfo.c (init_reg_sets_1, record_subregs_of_mode): Likewise.
> * reload1.c (maybe_fix_stack_asms, finish_spills): Likewise.
> * resource.c (find_dead_or_set_registers): Likewise.
> * sel-sched.c (mark_unavailable_hard_regs): Likewise.
FWIW, in case anyone is wondering, yes I see the operator overloads and
these patches seem like precisely the way we want to be using them.