From owner-freebsd-bugs Sun May 21 13:29:28 1995 Return-Path: bugs-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id NAA27710 for bugs-outgoing; Sun, 21 May 1995 13:29:28 -0700 Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id NAA27703 for ; Sun, 21 May 1995 13:29:21 -0700 Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.11/8.6.9) id NAA03831; Sun, 21 May 1995 13:29:05 -0700 From: "Rodney W. Grimes" Message-Id: <199505212029.NAA03831@gndrsh.aac.dev.com> Subject: Re: MAJOR problem with FreeBSD-2.0-RELEASE To: tege@cygnus.com (Torbjorn Granlund) Date: Sun, 21 May 1995 13:29:05 -0700 (PDT) Cc: freebsd-bugs@FreeBSD.org In-Reply-To: <199505212021.NAA25309@cygnus.com> from "Torbjorn Granlund" at May 21, 95 01:21:26 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 3320 Sender: bugs-owner@FreeBSD.org Precedence: bulk > > If this is the Intel Zappa board, make sure you are using TIN connectors > on the SIMM modules, if they are gold you may have a very serious memory > system problem. (The data book that came with your Zappa board will > say to use only Tin SIMM modules, believe it!!!). > > The SIMMs have tin connectors. Okay, good! > This really smells like memory system problems. Are you running a 90Mhz > or 100Mhz CPU?? The Zappa board is not rated for 100Mhz operation (or > at least the 2 board I looked at clearly stated 75 or 90Mhz but not > 100 Mhz.) > > I use a 90 MHz CPU. (Becasue of the 20ns SRAMs I didn't dare to set the > jumpers for a 66 MHz external clock rate.) That is why the board is rate for 90Mhz, and even then they have to be pushing the limits a little. All the other boards I use have 15nS cache chips, this is the first 90Mhz board I have seen with 20nS chips in it :-(. > > Another problem is the driver for the buslogic card (bt742.c?). It doesn't > > do synchronuous SCSI with my Buslogic KT-946C. I cannot get to > > ftp.cdrom.com, so I cannot check if the driver is improved. Is it? > > Humm.. been running bt946's in sync mode for over a year, what version > of the board/BIOS/firmware do you have? > > I don't remember, but the card is brand new. (I'll check the version in a > bit.) Brand new means little with Bus Logic, and the time it takes for new versions of the board to trickly through the distribution channel. I have found my old versions of the cards to actually work better then the newer versions!! [Just ask Jordan, he had a nightmare with the latest version of cards from Bus Logic, so have I, basically can't get the d*mn things to work with several of my newer motherboards, and due to lack of Bus Logic's technical support have shit canned the whole product line :-(] > Do you have the sync options enabled on the bt946C? > > They are enabled correctly. And the BIOS recognizes that the drive is > capable of synchronuous transfers! But then reports to FreeBSD that the board is in async mode, seems we had someone else with this problem, seems BL has changed a bit definition on us :-(. > > Note that we observed the exact same behaviour when using FreeBSD 2.0 on > > a completely different system (different CPU, motherboard, disks, SCSI > > card) . This makes it unlikely to be hardware problems. > > Ahh... FreeBSD 2.0, been a long time since I've run it, could very well > be problems in that release. Though I have run make world using it > to upgrade a system to 2.0 current back in February (initial boot strapping > of my build environment here after being gone for some time). That was > run on a Opti based P54C-90 with a BT946C and 16MB, would not have seen > a bounce buffer bug :-(. > > Could you possibly make a newer kernel that I can try? Could, but it will have lots of problems with 2.0 binaries, especially networking code. You could also grab the kernel from freefall, or grab the 950412 snap floppies and get a kernel from there. These will all have 2.0R compatibility problems, but at least it should be more stable. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Custom computers for FreeBSD