This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH]: Remove some unconditional SMT usage updates
- From: Daniel Berlin <dberlin at dberlin dot org>
- To: Diego Novillo <dnovillo at redhat dot com>
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Fri, 03 Mar 2006 12:10:44 -0500
- Subject: Re: [PATCH]: Remove some unconditional SMT usage updates
- References: <1141403785.29060.4.camel@localhost> <44087571.1050305@redhat.com>
On Fri, 2006-03-03 at 11:57 -0500, Diego Novillo wrote:
> On 03/03/06 11:36, Daniel Berlin wrote:
>
> > @@ -2194,6 +2196,7 @@ tree_sra (void)
> > scan_function ();
> > decide_instantiations ();
> > scalarize_function ();
> > + todoflags = TODO_update_smt_usage;
> >
> I would rather do this in either decide_instantiations. If any var in
> sra_candidates is an array, then flip TODO_update_smt_usage on todoflags
> (and make todoflags a file static).
>
Okay.
I mistakenly thought that if we get into this clause, we are definitely
going to scalarize, but i see that is not true.
> I don't think we SRA structures with arrays
I believe Richard Guenther added the ability to do this.
> , but if we do then the
> flipping could be done in the scalarize_* functions.
I'll make it flip it iff we scalarize things, which was my original
intent.
>
> The rest is fine.
>