This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH 061/236] combine.c: Use rtx_insn


On 08/06/14 11:20, David Malcolm wrote:
gcc/
	* combine.c (i2mod): Strengthen this variable from rtx to rtx_insn *.
	(struct reg_stat_struct): Likewise for fields "last_death", "last_set".
	(subst_insn): Likewise for this variable.
	(added_links_insn): Likewise.
	(struct insn_link): Likewise for field "insn".
	(alloc_insn_link): Likewise for param "insn".
	(struct undobuf): Likewise for field "other_insn".
	(find_single_use): Likewise for param "insn" and local "next".
	(combine_validate_cost): Likewise for params "i0", "i1", "i2", "i3".
	(delete_noop_moves): Likewise for locals "insn", "next".
	(create_log_links): Likewise for locals "insn", "use_insn".
	Strengthen local "next_use" from rtx * to rtx_insn **.
	(insn_a_feeds_b): Likewise for params "a", "b".
	(combine_instructions): Likewise for param "f" and locals "insn",
	"next", "prev", "first", "last_combined_insn", "link", "link1",
	"temp".  Replace use of NULL_RTX with NULL when referring to
	insns.
	(setup_incoming_promotions): Likewise for param "first"
	(set_nonzero_bits_and_sign_copies): Likewise for local "insn".
	(can_combine_p): Likewise for params "insn", "i3", "pred",
	"pred2", "succ", "succ2" and for local "p".
	(combinable_i3pat): Likewise for param "i3".
	(cant_combine_insn_p): Likewise for param "insn".
	(likely_spilled_retval_p): Likewise.
	(adjust_for_new_dest): Likewise.
	(update_cfg_for_uncondjump): Likewise, also for local "insn".
	(try_combine): Likewise for return type and for params "i3", "i2",
	"i1", "i0", "last_combined_insn", and for locals "insn",
	"cc_use_insn", "p", "first", "last", "i2_insn", "i1_insn",
	"i0_insn".  Eliminate local "tem" in favor of new locals
	"tem_note" and "tem_insn", the latter being an rtx_insn *.  Add a
	checked cast for now to rtx_insn * on the return type of
	gen_rtx_INSN.  Replace use of NULL_RTX with NULL when referring to
	insns.
	(find_split_point): Strengthen param "insn" from rtx to
	rtx_insn *.
	(simplify_set): Likewise for local "other_insn".
	(recog_for_combine): Likewise for param "insn".
	(record_value_for_reg): Likewise.
	(record_dead_and_set_regs_1): Likewise for local
	"record_dead_insn".
	(record_dead_and_set_regs): Likewise for param "insn".
	(record_promoted_value): Likewise.
	(check_promoted_subreg): Likewise.
	(get_last_value_validate): Likewise.
	(reg_dead_at_p): Likewise.
	(move_deaths): Likewise for param "to_insn".
	(distribute_notes): Likewise for params "from_insn", "i3", "i2"
	and locals "place", "place2", "cc0_setter".  Eliminate local "tem
	in favor of new locals "tem_note" and "tem_insn", the latter being
	an rtx_insn *.
	(distribute_links): Strengthen locals "place", "insn" from rtx to
	rtx_insn *.
---
OK.

As are patches #62-

Just to be clear, these are largely mechanical changes. Changes to signatures, variables and the like I'm just scanning without deep inspection. Changes in code get more attention, but they're usually NULL_RTX->NULL or checked casts. The checked casts I'm holding my nose in the hope that many are going away. I think after this kit does in, a scan of the remaining checked casts would be advisable :-)

jeff



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]