From owner-freebsd-current Tue Nov 25 17:40:54 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id RAA28469 for current-outgoing; Tue, 25 Nov 1997 17:40:54 -0800 (PST) (envelope-from owner-freebsd-current) Received: from nomis.Simon-Shapiro.ORG (nomis.i-Connect.Net [206.190.143.100]) by hub.freebsd.org (8.8.7/8.8.7) with SMTP id RAA28423 for ; Tue, 25 Nov 1997 17:40:44 -0800 (PST) (envelope-from shimon@nomis.Simon-Shapiro.ORG) Received: (qmail 519 invoked by uid 1000); 26 Nov 1997 01:40:48 -0000 Message-ID: X-Mailer: XFMail 1.2-beta-111797 [p0] on FreeBSD X-Priority: 3 (Normal) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit MIME-Version: 1.0 In-Reply-To: <199711252331.KAA00560@word.smith.net.au> Date: Tue, 25 Nov 1997 17:40:48 -0800 (PST) Reply-To: shimon@simon-shapiro.org Organization: Me, Just me... From: Simon Shapiro To: Mike Smith Subject: Re: Compiler Bug??? Cc: freebsd-current@FreeBSD.ORG, Nate Williams Sender: owner-freebsd-current@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk On 25-Nov-97 Mike Smith wrote: ... > Not exactly. I quote: > > ] Scenario 1: > ] > ] Xfmail of 17-Nov-97 with xforms 0.88: Compiles fine, then produces > ] dozens of ``Bad Attribute'' and `Bad Coordinates''. These come > from > ] libX11. Tearing my hair out did not help, so I boot UP (was on > SMP). > ] The problem is... GONE! > ] > ] Recompile on UP, reboot SMP... Problem is... Yup, GONE! > > This is actually: > > - code built while running SMP fails while running SMP. > - code built while running SMP works UP. > - code built while running UP works everywere. > > Either of our interpretations could be right. More data, Simon? OK, if you insist :-) It appears as if the SMP kernel I am running (from 10-Nov-97) has (wild guess!) some sort of a leak in the VM: * Innocent copiles crash with cc1 exiting on SEGFAULT. * Code copiled will crash in a strange way. * Fsck, upon reboot) will destroy (or try to) a file system because ALL the inodes need to be 208, not what they really are. These problems go away if I boot a uniprocessor kernel. The only constant is: If it was compiled under UP and run under UP it is good and stable. All other iterations and combinations have been observed. I have not upgraded to kernels past the 10th, as the SMP kernel boots fine, but instead of starting /etc/rcit just sits there in the default halt loop. I have received help on this but th ONLY thnig that really works is to start with the GENERIC kernel and move things around, deleting chaff, one step at a time, until a new config file is born. The new config file is materially identical to the old, but ordered a bit differently. Dunno why any of this makes any sense. But it does Simon BTW, different motherboard models react differently to these scenarios.