From owner-freebsd-questions Sun Jan 18 08:29:00 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id IAA20970 for questions-outgoing; Sun, 18 Jan 1998 08:29:00 -0800 (PST) (envelope-from owner-freebsd-questions@FreeBSD.ORG) Received: from mail.pernet.net (mail.pernet.net [205.229.0.40]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA20938 for ; Sun, 18 Jan 1998 08:28:45 -0800 (PST) (envelope-from neal@pernet.net) Received: from office.pernet.net (office.pernet.net [205.229.0.33]) by mail.pernet.net (8.8.5/8.8.4) with SMTP id KAA26524; Sun, 18 Jan 1998 10:37:21 -0600 (CST) Date: Sun, 18 Jan 1998 10:28:40 -0600 (CST) From: Neal Reply-To: neal@pernet.net To: Doug White cc: freebsd-questions@FreeBSD.ORG Subject: Re: double kernel faults In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk Nope. The second fault appears to happen when the kernel tries to sync the disks. I haven't traced ALL the way back, but it looks like this, time order being top is the start of the fault: actual fault happens trap() is called sync() is called msync() is called second fault happens trap() called sync() called msync() called reboots correctly. I don't have my notes handy(I'm in Houston for the weekend, yeah!), but I've found the actual location of the panic. Someplace along the line(inside of sync()), the vfs struct is dissapearing. When the actual sysctl macro gets called, on of the arguments is a pointer to a pointer from a NULL. I'll send a stack trace when I get back local. On Sat, 17 Jan 1998, Doug White wrote: > On Fri, 16 Jan 1998, Neal Rigney wrote: > > > > > I've got a news machine that continually page faults(I'm tracking the > > problem down right now), but here's the interesting question: > > > > When it page faults, it ALWAYS(with this error) faults a second time. In > > other words, I get "page fault while in kernel mode" etc(the REAL error I > > want to look at) immediately followed by another fault that of course > > scrolls all the information from the first fault off the screen. > > > > So, should I _really_ worry about the second fault, or chaulk it up to the > > first fault making the system go nuts? > > I'd be more interested in what the second fault is so we can resolve that > so we can actually *see* the details on the first fault ;-) > > Does the `scroll lock-up arrow' scrollback work during a panic? I forgot. > > Doug White | University of Oregon > Internet: dwhite@resnet.uoregon.edu | Residence Networking Assistant > http://gladstone.uoregon.edu/~dwhite | Computer Science Major > > > -- Neal Rigney, PERnet Communications, (409)729-4638 neal@mail.pernet.net