Date: Thu, 9 Oct 1997 22:55:33 +0200 From: Stefan Esser <se@FreeBSD.ORG> To: Philippe Regnauld <regnauld@deepo.prosa.dk> Cc: freebsd-scsi@FreeBSD.ORG Subject: Re: 2.2.2 anc NCR875 failures Message-ID: <19971009225533.39451@mi.uni-koeln.de> In-Reply-To: <19971009182853.57538@deepo.prosa.dk>; from Philippe Regnauld on Thu, Oct 09, 1997 at 06:28:53PM %2B0200 References: <19971008113725.46245@deepo.prosa.dk> <19971008225552.49139@mi.uni-koeln.de> <19971009182853.57538@deepo.prosa.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
On 1997-10-09 18:28 +0200, Philippe Regnauld <regnauld@deepo.prosa.dk> wrote: > Stefan Esser writes: > > > > No, your problem is different from the QUEUE FULL > > situation others are suffering from. But that may > > still hurt you, if you got revision LXY4 firmware > > in your Atlas II drives ... > > Oh great! Here's the info: > > The NCR chip says: > > 609-0392 410 > dp03283 > 9717r > > The FLASH rom says: > > Rev 2.00 FLASH 390F/U > > (this is a "Tekram 390" with ncr 53c875) Yes, that's the latest Tekram DC390F with Flash ROM ... > Output from dmesg with -v -v: (as you will see, I _have_ LXY4 > on both disks :-[ They may cause problems with tags at times. I do not know whether there is newer firmware, but it generally available from Quantums FTP site: ftp://ftp.qntm.com/pub/support/Firmware > FreeBSD 2.2.2-RELEASE #2: Thu Oct 9 18:21:09 CEST 1997 Well, you reminded me of the subject in a previous message, and I've got to admit that I just did not bother to read the full text of the subject line ... > CPU: AMD K6 (187.11-MHz 586-class CPU) What's the PCI bus clock ??? Anything beyond 33MHz may cause problems with some PCI chips. (The NCR controllers are known to work at 37.5MHz, but many other devices may cause spurious errors!) If you currently are running with 5 x 37.5, then you may want to try 6 x 30 instead. Even though your system will be some 10% slower, you may find that it becomes stable again ... > chip0 <generic PCI bridge (vendor=8086 device=7100 subclass=0)> rev 1 on pci0:0 > chip1 <generic PCI bridge (vendor=8086 device=7110 subclass=1)> rev 1 on pci0:1:0 > pci0:1:1: Intel Corporation, device=0x7111, class=storage (ide) [no driver assigned] > map(20): io(e000) > pci0:1:2: Intel Corporation, device=0x7112, class=0x0c, subclass=0x03 int d irq 9 [no driver assigned] > map(20): io(d800) > chip2 <generic PCI bridge (vendor=8086 device=7113 subclass=128)> rev 1 on pci0:1:3 Hmmm, have to make sure they will be correctly identified in 2.2.5 ... > vga0 <VGA-compatible display device> rev 0 int a irq 12 on pci0:10 What brand of VGA card is that ? > ncr0 <ncr 53c875 wide scsi> rev 3 int a irq 11 on pci0:12 > initial value of SCNTL3 = 03, final = 13 Ok. The register is set to a save value. This is the NCR driver before the Fast-20 patches went in. It uses the 875 with a 40MHz clock, just like a non-Ultra SCSI chip, and the value that surprised in me in your original problem report is in fact quite right for that driver revision. > (ncr0:0:0): "QUANTUM XP32275W LXY4" type 0 fixed SCSI 2 > sd0(ncr0:0:0): 20.0 MB/s (100 ns, offset 16) > (ncr0:1:0): "QUANTUM XP32275W LXY4" type 0 fixed SCSI 2 > sd1(ncr0:1:0): 20.0 MB/s (100 ns, offset 16) Ok. My advice: Make sure you are not running with a bus clock beyond that specified in the PCI spec. And with 37.5MHz, you are some 15% higher than allowed! I'm sure your system will become reliable, then. Regards, STefan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19971009225533.39451>