From owner-freebsd-questions Thu Jun 22 22:37: 8 2000 Delivered-To: freebsd-questions@freebsd.org Received: from firemoth.pkunk.net (firemoth.pkunk.net [63.201.19.138]) by hub.freebsd.org (Postfix) with ESMTP id 1300137B5E0 for ; Thu, 22 Jun 2000 22:36:54 -0700 (PDT) (envelope-from larry@pkunk.net) Received: from fury (fury.pkunk.net [63.201.19.139]) by firemoth.pkunk.net (8.9.3/8.9.3) with SMTP id WAA00841; Thu, 22 Jun 2000 22:36:21 -0700 (PDT) (envelope-from larry@pkunk.net) From: "Lawrence Cotnam Jr." To: Cc: Subject: Problems with SiS 5591 PCI IDE Date: Thu, 22 Jun 2000 22:36:20 -0700 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0) Importance: Normal X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6700 Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG I've been running FreeBSD 4.0 since its release on my Single Board computer system which employs a SiS 5591 PCI IDE controller. All has been just fine and happy until I recently upgraded the disk drive to a 13G drive capable of UDMA2. FreeBSD detects that the SiS 5591 is capable of UDMA2 and places the controller and drive in UDMA2 mode, at which point the system freezes when it attempts to mount the root filesystem. Efforts to correct the situation have resulted in many problems which has caused me to strongly consider reverting to FreeBSD 3.x due to all the problems I'm encountering. My first attempt to fix the problem was to recompile my kernel and making sure the option ATA_ENABLE_ATAPI_DMA was not defined. Rebooting this kernel still resulted in detection of UDMA2 capable devices and the freeze when attempting to mount the root file system. Second, needing my web and domain server back desperetely, I reverted to older WDC and WD drivers to boot the machine, disabling all ATA devices in my config file and rebuilding my kernel. This booted successfully, but after about 4 hours of operation, the driver reported a error during writing and disabled multi-sector I/O. No other adverse problems were detected initially. But worried I proceeded with more attempts to get the ATAPI drivers working again with my machine. Next, I looked into the source code for ata-dma.c, added a return just before it tries to probe for DMA capable devices and rebooted. This resulted in a usable system for a little while. After several hours of operation, I began to notice HARD WRITE errors occuring occasionally, as well as other indications of disk corruption occuring. Paniced, I reverted again to WD drivers without multi-sector I/O. So what's the deal? Is the IDE controller driver code just plain flaky now? I've noticed, when I boot with WD drivers, its even reporting a different controller text during device detection. With WD drivers I get "ide_pci0: ", when I boot with ATAPI drivers, I get "atapci0: " I want to stress that before upgrading the computer's disk drive, I've never had a single problem, so its definitely some sort of brokenness with the fact my new drive supports UDMA2. What I'd like is a reliable system, not UDMA support. Any suggestions in getting a reliable system with this mixture of hardware would be appreciated. Lawrence Cotnam Jr. phone: (775) 337-2536 email: larry@pkunk.net To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message