This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Your change to jump.c breaks conditional move?
- To: Richard Henderson <rth at cygnus dot com>
- Subject: Re: Your change to jump.c breaks conditional move?
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Fri, 05 Jun 1998 00:22:45 -0600
- cc: "H.J. Lu" <hjl at lucon dot org>, egcs-patches at cygnus dot com, egcs at cygnus dot com
- Reply-To: law at cygnus dot com
In message <19980604115358.A4098@dot.cygnus.com>you write:
> On Thu, Jun 04, 1998 at 11:27:08AM -0700, H.J. Lu wrote:
> > I think either way we should handle it. We can put abort () there if
> > we don't think it is very likely. Otherwise, we handle it like PLUS.
>
> Ok, here is an alternate patch. I also realized I should have
> been checking the other side of the op for a symbol.
Actually, I don't think the check for SYMBOL_REF is sufficient.
The address could be a SYMBOL_REF, (CONST (PLUS (SYMBOL_REF) (CONST_INT)),
LABEL_REF or even a CONST_INT.
I'm also not sure checking for PLUS/MINUS is enough either since
some PIC implementations use LO_SUM instead of PLUS to add the
pic register to the lower bits of a symbolic address.
jeff