This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: tree-ssa branch broken
- From: Daniel Berlin <dberlin at dberlin dot org>
- To: Diego Novillo <dnovillo at redhat dot com>
- Cc: Andreas Jaeger <aj at suse dot de>, <gcc-bugs at gcc dot gnu dot org>
- Date: Thu, 27 Jun 2002 16:59:29 -0400 (EDT)
- Subject: Re: tree-ssa branch broken
On Thu, 27 Jun 2002, Diego Novillo wrote:
> On Thu, 27 Jun 2002, Andreas Jaeger wrote:
>
> >
> > Current CVS of teh tree-ssa branch does not bootstrap for me on
> > i686-linux-gnu, xgcc is miscompiled:
> >
> Very strange. The overnight SPEC95 tester didn't have a problem
> with it. I'm wondering if this is the first hint that I should
> actually mark basic_block and the tree SSA data structures for
> garbage collection. Right now we've been ignoring it, because
> it's rather annoying to fix.
No, it's not.
I ran through the code paths, and ggc_collect is never called in
between the start of simplification, and the end of optimize_tree.
Thus, the memory can't be disappearing on you when you are still
trying to use it.
The tree annotations also are always created at the beginning, regardless
of if TREE_ANN is null or not (IE the thing initialially creating refs
calls create_tree_ann, rather than get_tree_ann).
I'll check on BB_ANN's. If we aren't always forcing creation of them for
the the first thing to touch them, that could be the bug (since that
field might not be clear, but really garbage.)
> Diego. > >