This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [Patches] Code cleanup - remove unused variables - part 2
- From: Richard Henderson <rth at redhat dot com>
- To: Reichelt <reichelt at igpm dot rwth-aachen dot de>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Wed, 11 Sep 2002 14:55:07 -0700
- Subject: Re: [Patches] Code cleanup - remove unused variables - part 2
- References: <200209111609.SAA27698@numa6.igpm.rwth-aachen.de>
On Wed, Sep 11, 2002 at 06:09:23PM +0200, Reichelt wrote:
> +++ gcc-new/gcc/genattrtab.c Fri Sep 6 13:03:10 2002
> @@ -2562,11 +2562,10 @@ simplify_cond (exp, insn_code, insn_inde
> int len = XVECLEN (exp, 0);
> rtx *tests = (rtx *) xmalloc (len * sizeof (rtx));
> int allsame = 1;
> - char *first_spacer;
> rtx ret;
>
> /* This lets us free all storage allocated below, if appropriate. */
> - first_spacer = (char *) obstack_finish (rtl_obstack);
> + obstack_finish (rtl_obstack);
Ok, I've looked at this further. All references to obstacks
in genattrtab.c are now garbage and can be removed.
> +++ gcc-new/gcc/jump.c Fri Sep 6 13:57:28 2002
> @@ -78,10 +78,8 @@ rebuild_jump_labels (f)
> rtx f;
> {
> rtx insn;
> - int max_uid = 0;
> -
> - max_uid = init_label_info (f) + 1;
>
> + init_label_info (f);
Incidentally, this now means that init_label_info can return void,
and thus the collection of largest_uid in that function is garbage.
> +++ gcc-new/gcc/loop.c Fri Sep 6 14:11:42 2002
> @@ -5915,10 +5905,7 @@ check_final_value (loop, v)
> struct induction *v;
> {
> struct loop_ivs *ivs = LOOP_IVS (loop);
> - struct iv_class *bl;
> rtx final_value = 0;
> -
> - bl = REG_IV_CLASS (ivs, REGNO (v->src_reg));
ivs is now garbage too.
> @@ -819,9 +817,6 @@ process_note_predictions (bb, heads, dom
> delete_insn (insn);
> }
> }
> - for (e = bb->succ; e; e = e->succ_next)
> - if (!(e->flags & EDGE_FAKE))
> - noreturn_block = 0;
> if (contained_noreturn_call)
This one looks like a legitimate error. That IF should
be testing noreturn_block not contained_noreturn_call.
> @@ -2595,10 +2594,8 @@ want_to_remat (x)
> SET_SRC (PATTERN (remat_test_insn)) = x;
> /* XXX For now we don't allow any clobbers to be added, not just no
> hardreg clobbers. */
> - return ((icode = recog (PATTERN (remat_test_insn), remat_test_insn,
> - &num_clobbers)) >= 0
> - && (num_clobbers == 0
> - /*|| ! added_clobbers_hard_reg_p (icode)*/));
> + return (recog (PATTERN (remat_test_insn), remat_test_insn, &num_clobbers) >= 0
> + && num_clobbers == 0);
> }
This one looks like it should be left alone for the time being.
If you like, to remove the warning from your radar, rearrange as
icode = recog (PATTERN (remat_test_insn), remat_test_insn, &num_clobbers);
return (icode >= 0 && ...)
I've committed everything except these last two hunks.
r~