This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [RFA] Fix PR rtl-optimization/30807
- From: Kaz Kojima <kkojima at rr dot iij4u dot or dot jp>
- To: ebotcazou at adacore dot com
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Mon, 06 Jul 2009 22:30:32 +0900 (JST)
- Subject: Re: [RFA] Fix PR rtl-optimization/30807
- References: <20090611.065131.04692084.kkojima@rr.iij4u.or.jp> <200907041542.20579.ebotcazou@adacore.com>
Eric Botcazou <ebotcazou@adacore.com> wrote:
> The assertion seems superfluous to me, operand 1 of REG_SUM is always BASE.
>
> So I'd just write:
>
> + /* For every new use of REG_SUM, we have to record the use
> + of BASE therein, i.e. operand 1. */
> + for (i = reg_state[regno].use_index;
> + i < RELOAD_COMBINE_MAX_USES; i++)
> + reload_combine_note_use
> + (&XEXP (*reg_state[regno].reg_use[i].usep, 1),
> + reg_state[regno].reg_use[i].insn);
>
> OK with this change if it still passes a full testing cycle.
Thanks for reviewing. The patch with that change passed the full
tests on i686-pc-linux-gnu and powerpc-apple-darwin9.5.0.
Applied as revision 149282.
Regards,
kaz