From owner-freebsd-bugs Sun May 21 13:23:45 1995 Return-Path: bugs-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id NAA27519 for bugs-outgoing; Sun, 21 May 1995 13:23:45 -0700 Received: from hda.com (hda.com [199.232.40.182]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id NAA27508 for ; Sun, 21 May 1995 13:23:38 -0700 Received: (dufault@localhost) by hda.com (8.6.9/8.3) id QAA18523; Sun, 21 May 1995 16:23:28 -0400 From: Peter Dufault Message-Id: <199505212023.QAA18523@hda.com> Subject: Re: MAJOR problem with FreeBSD-2.0-RELEASE To: sef@kithrup.com (Sean Eric Fagan) Date: Sun, 21 May 1995 16:23:27 -0400 (EDT) Cc: bde@zeta.org.au, davidg@Root.COM, freebsd-bugs@FreeBSD.org, tege@cygnus.com In-Reply-To: <199505211847.LAA13914@kithrup.com> from "Sean Eric Fagan" at May 21, 95 11:47:44 am X-Mailer: ELM [version 2.4 PL24] Content-Type: text Content-Length: 2262 Sender: bugs-owner@FreeBSD.org Precedence: bulk Sean Eric Fagan writes: > > >There have been several reports that bounce buffering is broken. The bt > >driver always enables bounce buffering, although this at best wastes time > >on non-ISA systems. > > Some more information: > > the system has a Quantum 4.1GByte drive on it (one of the Grand Prix, > supposedly, although it apparantly identifies itself as a slightly different > model). > > I just tried running a kernel without bounce buffers, without an ed0 driver, > and with MAXMEM set to 16384. I booted, and then tried to build bash by > > cd /usr/src/local/bash* > make > > It first gave a warning/error message from the buslogic driver, saying > something about there being more than 33 mbxs (I think it was); there was > then a message from the scsi code saying "oops" something wasn't queued. > The system then got a kernel page fault, and paniced. > > Before that, with slightly different kernels, I'd noticed that > > cd /sys/i386/conf > vi TEGE > > would get the buslogic warnings/errors, and vi would come up with an empty > file, or core dump. If, however, I did > > vi > :e TEGE > > it would work. > > I think he is going to install Linux. This is the second motherboard he's > been unsuccessful in running FreeBSD-2.0 on. > > Sean. > "oops not queued" should be changed to be a panic (kick me if you disagree, Julian, but I'd change it now even before 2.0.5 burns). Now I'll go out on a limb and guess: our Buslogic support is broken. Someone has to provide an alternate SCSI board to these customers until it is fixed. In spite of what Rodney said earlier, I like Buslogic because they have been extremely helpful to me in the past, and so the help they gave me cancels the four hour phone wait Rodney had, and FreeBSD customers should give them another chance and let's not put them on our "Don't buy this stuff" list (It wasn't that long ago that I regularly saw "buy Buslogic" postings). Is anyone in a position to get a known working PCI SCSI board to sef? This is a FreeBSD support challenge and I'd like to see us rise to it. Peter -- Peter Dufault Real Time Machine Control and Simulation HD Associates, Inc. Voice: 508 433 6936 dufault@hda.com Fax: 508 433 5267