From owner-freebsd-alpha Thu Mar 30 12:59:21 2000 Delivered-To: freebsd-alpha@freebsd.org Received: from uni4nn.gn.iaf.nl (osmium.gn.iaf.nl [193.67.144.12]) by hub.freebsd.org (Postfix) with ESMTP id E7E6837B8D9; Thu, 30 Mar 2000 12:59:15 -0800 (PST) (envelope-from wilko@yedi.iaf.nl) Received: from yedi.iaf.nl (uucp@localhost) by uni4nn.gn.iaf.nl (8.9.2/8.9.2) with UUCP id WAA21479; Thu, 30 Mar 2000 22:53:07 +0200 (MET DST) Received: (from wilko@localhost) by yedi.iaf.nl (8.9.3/8.9.3) id WAA04115; Thu, 30 Mar 2000 22:52:48 +0200 (CEST) (envelope-from wilko) Date: Thu, 30 Mar 2000 22:52:47 +0200 From: Wilko Bulte To: Andrew Gallatin Cc: wilko@freebsd.org, Dave Haney , freebsd-alpha@freebsd.org Subject: Re: Unexpected machine check Message-ID: <20000330225247.C3785@yedi.iaf.nl> Reply-To: wilko@freebsd.org References: <20000330201513.A1750@yedi.iaf.nl> <14563.40742.553401.107502@grasshopper.cs.duke.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0.1i In-Reply-To: <14563.40742.553401.107502@grasshopper.cs.duke.edu>; from gallatin@cs.duke.edu on Thu, Mar 30, 2000 at 01:47:41PM -0500 X-OS: FreeBSD 3.4-STABLE X-PGP: finger wilko@freebsd.org Sender: owner-freebsd-alpha@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Thu, Mar 30, 2000 at 01:47:41PM -0500, Andrew Gallatin wrote: > Wilko Bulte writes: > > > > There is definitely a bogon somewhere as far as X on Alpha goes. The problem > > seems to be core logic chipset related, people fortunate enough to own > > Tsunami based systems do not seem to have these problems. Also the type of > > VGA chip seems to matter. > > I suspect the real problem is bugs in the X server & that's why some > cards (Matrox, 3DLabs Permedia) work and others don't. The Tsunami > chipset itself just masks bugs. Like a PC, writes to invalid I/O > ports, etc are ignored & reads return -1. Such accesses will cause > machine checks on all other alphas. > > If somebody was to go through the effort of building one of these > failing X servers with symbols, the PC mentioned in the machine check I'll try doing that next week. Is there anything special to do when building the port? > would likely be the PC of the offending instruction. Or close to it. Would that be back-traceable to if it is the X server or the kernel? > In fact, I wonder if we couldn't hack up a kernel to send a sigbus to > the offending X server & get a core dump rather than panicing. Sounds reasonable.. a panic is a bit drastic ;) -- Wilko Bulte Arnhem, The Netherlands http://www.tcja.nl The FreeBSD Project: http://www.freebsd.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message