This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] Fix make_goto_expr_edges for computed gotos
- From: law at redhat dot com
- To: Diego Novillo <dnovillo at redhat dot com>
- Cc: Zdenek Dvorak <rakdver at atrey dot karlin dot mff dot cuni dot cz>, "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>
- Date: Thu, 21 Aug 2003 22:31:35 -0600
- Subject: Re: [tree-ssa] Fix make_goto_expr_edges for computed gotos
- Reply-to: law at redhat dot com
In message <1061395471.3380.236.camel@frodo.toronto.redhat.com>, Diego Novillo
writes:
>Do you have a test case? I'm thinking that it may be better to have
>FORCED_LABELs and NONLOCAL_LABELs start their own basic block. Mixing
>them with regular labels penalizes regular labels with too many edges
>that are not destined to them. In which case, you would need to change
>stmt_starts_bb_p.
>
>However, I don't have a strong preference. What do others think?
I've got no strong opinions either. One possibility would be to only
collapse them to a single block if doing so would not change the
in-degree of the block with the normal label (or maybe only increases
the in-degree by a small amount).
Regardles, I doubt it's that big of an issue from a performance
standpoint.
Jeff