From owner-freebsd-stable Sun Sep 2 14:30:59 2001 Delivered-To: freebsd-stable@freebsd.org Received: from lerami.lerctr.org (lerami.lerctr.org [207.158.72.11]) by hub.freebsd.org (Postfix) with ESMTP id AC9D137B401; Sun, 2 Sep 2001 14:30:50 -0700 (PDT) Received: from lerami.lerctr.org (localhost [127.0.0.1]) by lerami.lerctr.org (8.12.0.Beta19/8.12.0.Beta19/20010730/$Revision: 1.25 $) with ESMTP id f82LUmXl027387; Sun, 2 Sep 2001 16:30:48 -0500 (CDT) Received: (from ler@localhost) by lerami.lerctr.org (8.12.0.Beta19/8.12.0.Beta19/Submit) id f82LUlXr027386; Sun, 2 Sep 2001 21:30:47 GMT Date: Sun, 2 Sep 2001 16:30:47 -0500 From: Larry Rosenman To: Marcel Moolenaar Cc: John Baldwin , freebsd-stable@FreeBSD.org, qa@FreeBSD.org, Eric Masson Subject: Re: cputype=486 Message-ID: <20010902163046.A26933@lerami.lerctr.org> References: <20010901114903.D11062@athlon.pn.xcllnt.net> <20010901161054.B13047@athlon.pn.xcllnt.net> <20010902084832.B2510@lerami.lerctr.org> <20010902111131.B478@dhcp01.pn.xcllnt.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20010902111131.B478@dhcp01.pn.xcllnt.net> User-Agent: Mutt/1.3.22.1i X-Mailer: Mutt http://www.mutt.org/ Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG * Marcel Moolenaar [010902 13:11]: > On Sun, Sep 02, 2001 at 08:48:32AM -0500, Larry Rosenman wrote: > > > Ah yes. In that case it only works if you link shared, but then you > > > have different problems. I guess this rules out CPUTYPE as a generic > > > tunable. If you want the highest possible performance, you give up > > > on portibility. You can't have it both... > > Is this a *NO*, we're not interested in fixing it? > > It isn't as simple as that. The problem is hairy. This means that > it takes time to come up with a solution that is at least acceptable > and it also takes time implementing it. Since a fix is not crucial > to the proper functioning of FreeBSD, it will not compete for the > highest slots on people's backlog. So it's a very *LOW* priority, and won't be fixed soon, if ever. > > > If so, some BIG notices should be in /etc/defaults/make.conf around > > the CPUTYPE setting. > > What, besides the notice that setting CPUTYPE in /etc/make.conf is > the worst of all possible places, are thinking about? that if you are cross-building, CPUTYPE shouldn't be set, that CPUTYPE being set to higher than the lowest processor you are building for *WILL* break (give my example of i486 build on a P3 world if you want), Or, remove the CPUTYPE option altogether until it can be made cross-build safe. I was *VERY* surprised that a cross-build picked up *ANY* of the host libraries for stuff running on the target. My logical brain ass/u/med that a cross-build was totally safe from the libraries on the host box. Obviously, I can't force any of this. Should I file a PR so it won't get lost? I, obviously, have removed CPUTYPE from my make.conf for this setup. LER -- Larry Rosenman http://www.lerctr.org/~ler Phone: +1 972-414-9812 E-Mail: ler@lerctr.org US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message