This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH 3/5] New intraprocedural Scalar Reduction of Aggregates.
- From: Richard Guenther <rguenther at suse dot de>
- To: Martin Jambor <mjambor at suse dot cz>
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>, Jan Hubicka <hubicka at ucw dot cz>
- Date: Wed, 29 Apr 2009 12:27:10 +0200 (CEST)
- Subject: Re: [PATCH 3/5] New intraprocedural Scalar Reduction of Aggregates.
- References: <20090428100429.051912011@virgil.suse.cz> <20090428100449.910685488@virgil.suse.cz>
On Tue, 28 Apr 2009, Martin Jambor wrote:
> This is the new intraprocedural SRA. I have stripped off the
> interprocedural part and will propose to commit it separately later.
> I have tried to remove almost every trace of IPA-SRA, however, two
> provisions for it have remained in the patch. First, an enumeration
> (rather than a boolean) is used to distuinguish between "early" and
> "late" SRA so that other SRA modes can be added later on. Second,
> scan_function() has a hook parameter and a void pointer parameter
> which are not used in this patch but will be by IPA-SRA.
>
> Otherwise, the patch is hopefully self-contained and the bases of its
> operation is described by the initial comment.
>
> The patch bootstraps (on x86_64-linux-gnu but I am about to try it on
> hppa-linux-gnu too) but produces a small number of testsuite failures
> which are handled by the two following patches.
>
> Thanks,
>
> Martin
>
>
> 2009-04-27 Martin Jambor <mjambor@suse.cz>
>
> * tree-sra.c (enum sra_mode): The whole contents of the file was
> replaced.
That looks odd ;) I would make it
* tree-sra.c: New implementation of SRA.
and in the usual way list any exported functions that were removed and
list new exported functions that were added.
Can you instead of posting a diff post the new contents? That would
be way easier to review.
Thanks,
Richard.