This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Ongoing problems with resource.c changes
- To: Richard Henderson <rth at cygnus dot com>
- Subject: Re: Ongoing problems with resource.c changes
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Mon, 01 Nov 1999 23:55:17 -0700
- cc: gcc at gcc dot gnu dot org
- Reply-To: law at cygnus dot com
In message <19991101224624.A27869@cygnus.com>you write:
> On Mon, Nov 01, 1999 at 09:20:18PM -0700, Jeffrey A Law wrote:
> > The problem is reorg calls redirect_jump. If in the process of redirecti
> ng
> > a jump we may delete code. Both actions (redirecting the jump & deleting
> > code) can make the cfg data inaccurate.
>
> Oh shit, we're screwed.
I concur.
> Any suggestions? I don't wanna rewrite reorg to do the right
> thing just now -- I just want to get the tree stable again.
Agreed -- I don't think we want to rewrite jump and/or reorg at the
moment.
> I could back out the changes from the last day or three and try
> and figure out why we discovered the wrong block... Or just
> disable peephole2 entirely for the nonce.
I'd say let's roll back and try to figure out why we got the wrong
block # for peep2.
Another approach would be to use different routines to map from an
insn to its basic block until we can rewrite jump & reorg to be
cfg-friendly.
jeff