This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] folding fix 2
- From: Jan Hubicka <jh at suse dot cz>
- To: law at redhat dot com
- Cc: Jan Hubicka <jh at suse dot cz>, gcc-patches at gcc dot gnu dot org,dnovillo at redhat dot com, rth at redhat dot com
- Date: Tue, 6 Jan 2004 16:57:51 +0100
- Subject: Re: [tree-ssa] folding fix 2
- References: <20031221101036.GE17542@kam.mff.cuni.cz> <200401060544.i065iUxH011481@speedy.slc.redhat.com>
> In message <20031221101036.GE17542@kam.mff.cuni.cz>, Jan Hubicka writes:
> >Hi,
> >fold folds a==b into different copy of a==b that waste memory and can
> >cause some optimizations thesting whether something simplified to loop.
> >I also have sanity checking veirfying that statements don't fold that
> >needs this.
> >OK?
> >Bootstrapped/regteste i686-pc-gnu-linux.
> >
> >Honza
> >
> >2003-12-21 Jan Hubicka <jh@suse.cz>
> > * fold-const.c (fold): Do not rebuild comparison when nothing
> > changed.
> This is fine. Please install this patch.
Thanks. The last fix
(http://gcc.gnu.org/ml/gcc-patches/2003-12/msg01846.html)
is last step preventing me from folding each statement after
gimplification solving some problems with non-canonical address
expressions confusing alias analysis of arrays. With that I can get
basic of SRA working on that and hopefully remove addressof RTL
construct (in my tests doing this right now brings some regressions all
I tracked down were related to small structures containing arrays or
nested substructures)
Honza
>
> Thanks,
> jeff
>
>