Re: CPUTYPE considered harmful? (was: Lot's of SIGILL, SIGSEGV)

From: Andre Guibert de Bruet <andy_at_siliconlandmark.com>
Date: Tue, 19 Aug 2003 11:28:05 -0400 (EDT)
On Tue, 19 Aug 2003, Stefan Bethke wrote:

> On Montag, 18. August 2003 23:15 Uhr +0200 Pawel Jakub Dawidek
> <nick_at_garage.freebsd.pl> wrote:
>
> > On Sun, Aug 17, 2003 at 08:00:54PM -0700, David O'Brien wrote:
> > +> > This is a FAQ. In the future, please search the archives before
> > posting. +> >
> > +> > At this moment in time, 'p4' isn't a safe CPUTYPE (It produces broken
> > +> > code). 'p3' or 'i686' are what's recommended for Pentium 4s.
> > +>
> > +> Andre, I think you are out of date -- CPUTYPE=p4 is now safe with GCC
> > +> 3.3.1.
> >
> > I think he is right, because when upgrading host where was gcc3.2 to
> > current -CURRENT (with gcc3.3) 'make world' builds make(1) in first
> > place and it is builded by gcc3.2 with CPUTYPE=p4, so it will be broken.
> >
> > So gcc have to be upgraded in first place (with CPUTYPE=p3).
>
> Hhm, sounds reasonable.
>
> However, I had the exact same problem updating from a 5.1-RC to a recent
> current on a P III 900, and there, I had CPUTYPE=p3 in make.conf.
>
> Removing CPUTYPE eventually gave me back working systems (I did restore
> 5.1-R bits prior to make world). Unfortunatly, I don't have the resources
> to investigate this further, but for the time being, I will not use CPUTYPE
> until others can confirm it's safe :-)

I've been running my laptop with a world that was built with CPUTYPE=p3
(GCC 3.3.1) for close to a month. I haven't noticed anything odd yet.

Regards,

> Andre Guibert de Bruet | Enterprise Software Consultant >
> Silicon Landmark, LLC. | http://siliconlandmark.com/    >
Received on Tue Aug 19 2003 - 06:28:07 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:19 UTC