This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: long pending patch for tree-profiling-branch
- From: Olga Golovanevsky <OLGA at il dot ibm dot com>
- To: Daniel Berlin <dberlin at dberlin dot org>
- Cc: Daniel Berlin <dberlin at us dot ibm dot com>, dnovillo at redhat dot com, gcc-patches at gcc dot gnu dot org, Ira Rosen <IRAR at il dot ibm dot com>, jh at suse dot cz, Steven Bosscher <stevenb at novell dot com>
- Date: Mon, 25 Jul 2005 14:37:57 +0300
- Subject: Re: long pending patch for tree-profiling-branch
Daniel Berlin <dberlin@dberlin.org> wrote on 24/07/2005 19:33:10:
> > * tree-vect-analyze.c: (ipa-prop.h): New include file.
> > (vect_base_addr_differ_p): User ipaa interface to
> > disambiguate formals.
>
> You really shouldn't be doing that there.
>
> I assume you are only doing it because may_alias_p isn't doing what it
> should?
>
Actually yes. I did it to show that ipaa info is kept and can be used.
If there was oracle (may_alias_p, or any one else) into which this
info can be integrated, I would use it. But for now such oracle does
not exist and when it does I will gladly remove this code.
By the way, for now the vectorizer does not use may_alias_p
(the latest code in autovect branch). It actually use type_mem_tag and
is_aliased_with function in which may_aliases field is used.
Do you think it will stay so, and ipaa info will be
incorporated in may_aliases/type_mem_tag, or alternatively, it will
be oracle like may_alias_p that will be called directly by vectorizer?