This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch] Fix SPEC slowdown from function specific branch
- From: Michael Matz <matz at suse dot de>
- To: "H.J. Lu" <hjl dot tools at gmail dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Thu, 8 Apr 2010 15:43:00 +0200 (CEST)
- Subject: Re: [patch] Fix SPEC slowdown from function specific branch
- References: <Pine.LNX.4.64.0808081611430.6093@wotan.suse.de> <o2x6dc9ffc81004071137te8ddd7a7v209c07e2cbfdeef@mail.gmail.com>
Hi,
On Wed, 7 Apr 2010, H.J. Lu wrote:
> > I have no real idea why it had to be moved into optimization_options
> > initially but it's certainly the wrong place. ÂThis fixes the severe SPEC
> > regression from r138154 .
> >
> > regstrapping in progress on i686 and x86_64. ÂOkay for trunk?
> >
> >
> > Ciao,
> > Michael.
> > --
> > Â Â Â Â* i386/i386.c: (override_options): Move initialisation from
> > Â Â Â Âflag_schedule_insns_after_reload to here from ...
> > Â Â Â Â(optimization_options): ... here.
> >
>
> This caused:
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43671
As written in the bug report, r153685 still worked, the patch above was
r138953. And if anything it just exposed a latent bug with
-fsched2-use-superblocks .
Ciao,
Michael.