This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Scheduler questions (related to PR17808)
- From: Steven Bosscher <stevenb at novell dot com>
- To: Vladimir Makarov <vmakarov at redhat dot com>
- Cc: Andrey Belevantsev <abel at ispras dot ru>, rearnsha at gcc dot gnu dot org, gcc at gcc dot gnu dot org
- Date: Tue, 5 Jul 2005 20:57:26 +0200
- Subject: Re: Scheduler questions (related to PR17808)
- References: <200506300046.22545.stevenb@suse.de> <42C3FF51.9040607@ispras.ru> <42CAD1CB.8050100@redhat.com>
On Tuesday 05 July 2005 20:30, Vladimir Makarov wrote:
> Andrey Belevantsev wrote:
> > Vladimir Makarov wrote:
> >> I'll look at this PR today.
> >
> > We've looked today at this issue. We think the problem is that
> > proposed patch of sched_get_condition() treats conditional jumps
> > likely to COND_EXECs, but it doesn't fix other places in sched-deps,
> > where COND_EXECs are considered. Maxim Kuvyrkov proposed the attached
> > patch, which allows gcc to bootstrap on ia64 and fixes the testcase in
> > PR.
>
> This patch is ok to fix PR17808. You just need to remove conditional
> #if and Richard's comment in sched_get_condition.
Don't do that just yet.
Vlad, you have missed that PR17808 is in fact two separate bugs. That
is in part my mistake, I should have opened a new PR for the missing
dependencies on ia64. But I thought it was obvious that we really have
two separate bugs here.
There is no code in sched-rgn to keep the jump at the end in the case
that the COND_EXEC_COND of a cond_exec insn and the jump condition of a
jump_insn are mutually exclusive. My patch that you just dismissed as
not necessary pretty clearly explained that, and I'm not sure why you
just ignored my explanation.
If you remove that #if 0 from sched_get_condition, you will break all
cond_exec targets except ia64.
Gr.
Steven