Re: PING: [patch] minor typo in modulo-sched.c

[Ayal Zaks <ZAKS at il dot ibm dot com>]

> Hello, Ayal and Revital,
>
> Could you have a look at this simple patch? Thanks.
>

Sorry for the delay, thanks for the reminder.

The patch is ok. It would be good to add an assert for stage_count >= 1;
could you please test that, while you're at it?

Thanks, Ayal.


> Bingfeng Mei
>
> Broadcom UK
>
>
> > -----Original Message-----
> > From: gcc-patches-owner@gcc.gnu.org
> > [mailto:gcc-patches-owner@gcc.gnu.org] On Behalf Of Bingfeng Mei
> > Sent: 07 July 2009 16:52
> > To: gcc-patches
> > Cc: Revital1 Eres; Ayal Zaks
> > Subject: [patch] minor typo in modulo-sched.c
> >
> > Hello,
> > I believe from how stage_count is calculated
> >
> > #define PS_STAGE_COUNT(ps) ((PS_MAX_CYCLE (ps) - PS_MIN_CYCLE (ps) \
> >               + 1 + (ps)->ii - 1) / (ps)->ii)
> >
> >
> > it is always equal or bigger than 1. So the following check
> > in the code
> > is not correct. OK for trunk?
> >
> >
> > Cheers,
> > Bingfeng Mei
> > Broadcom UK
> >
> >
> > 2009-07-07  Bingfeng Mei  <bmei@broadcom.com>
> >
> >          * modulo-sched.c (sms_schedule): stage_count <= 1 as
> > correct comparison
> >
> > Index: modulo-sched.c
> > ===================================================================
> > --- modulo-sched.c      (revision 149334)
> > +++ modulo-sched.c      (working copy)
> > @@ -1161,7 +1161,7 @@
> >
> >        /* Stage count of 1 means that there is no interleaving between
> >           iterations, let the scheduling passes do the job.  */
> > -      if (stage_count < 1
> > +      if (stage_count <= 1
> >           || (count_init && (loop_count <= stage_count))
> >           || (flag_branch_probabilities && (trip_count <=
> > stage_count)))
> >         {
> >
> >