This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: -mpower4 sched2 bug miscompiles linux kernel
- From: Alan Modra <amodra at bigpond dot net dot au>
- To: David Edelsohn <dje at watson dot ibm dot com>
- Cc: Geoff Keating <geoffk at geoffk dot org>, gcc-patches at gcc dot gnu dot org
- Date: Tue, 10 Sep 2002 11:40:46 +0930
- Subject: Re: -mpower4 sched2 bug miscompiles linux kernel
- References: <amodra@bigpond.net.au> <200209100153.VAA28060@makai.watson.ibm.com>
On Mon, Sep 09, 2002 at 09:53:22PM -0400, David Edelsohn wrote:
> This patch generally looks like the right direction to clean up
> that function, but the function either should use the _nv forms of
> adjust_address or the non-_nv forms. I think it should always use the
> non-_nv forms, but I'm not sure. I do not see any reason that it should
> differ between the STRING and non-STRING instructions.
The reason I used adjust_automodify_address_nv in the STRING case is
that I figured we'll always be generating a valid address, since we
construct the address rtx via copy_addr_to_reg. That's not true now
for non-STRING insns, so we can't use the _nv variant there. As far
as I can see, it shouldn't hurt to use the non _nv variant in both
places, so I'm happy to modify the patch to do that.
--
Alan Modra
IBM OzLabs - Linux Technology Centre