Date: Sun, 23 Feb 1997 22:11:16 -0800 From: jehamby@lightside.com (Jake Hamby) To: hackers@freebsd.org Subject: Help with nasty BusLogic SCSI card! Message-ID: <199702240611.WAA01315@lightside.com>
next in thread | raw e-mail | index | archive | help
This is not strictly a FreeBSD question, but involves a peculiar hardware configuration which seems to thwart any 32-bit OS. There is a Micron PCI Powerstation (90MHz Pentium) at work with a BusLogic BT-946C SCSI controller. I seem to remember that it is an older revision of the card. Anyway, we have had big problems configuring the card before. Once it is configured, we've had no problems with it, but last Friday I wanted to attempt to upgrade this Exabyte 8mm tape drive's firmware, which uses a DOS program. I was unsuccessful running the program on another PC, and I thought I might try the other system because it has a diferent SCSI controller. Anyway, I made a DOS boot disk (the system was running NT before), and the SCSI driver was searching the wrong port. Rather than fix the CONFIG.SYS, I thought to change the port of the SCSI card in the BusLogic BIOS... big mistake (which I should've known better considering the history of the PC!). Choosing Auto Config resetted the card back to settings which _should_ work but the card refuses to work now! The way in which the card doesn't work is peculiar, as well: It will boot an OS (for example NT, or the FreeBSD boot floppy), and correctly probe for devices, but once it starts to read from the hard drive (i.e. at sysinstall in FreeBSD, or the blue screen in Windows NT) it hangs. Under NT, it seems to repeatedly probe the disks, as the hard drive light will flash every few seconds, and eventually it will blue-screen crash (with the BUSLOGIC.SYS driver mentioned prominently in the crash). I remember that previously, in order to get it to work, we had to change the PCI Interrupt pin (and possibly set it to Edge trigger). But this time, I tried all eight combinations (A, B, C, D, with Level or Edge-trigger), and none of them worked. The PCI guru helping us correctly noted that there should be no reason to change the interrupt if there are no other bus-mastering cards in the system (the only other PCI card is a Matrox Ultima), but that's what needed to be done before. Also, like an idiot, I never wrote down the correct settings for all of the Buslogic parameters, and the "Auto Config" menu tends to reset just about everything. D'oh! Also, I know that this particular motherboard has a Flash BIOS, and I had upgraded it to a newer "Plug-and-Play" BIOS early last year. When I did that, I remember that I had to change the interrupt pin yet again. I will try on Monday down-grading to the older BIOS, but I still don't understand why this might be a factor. Anyway, since I know there are PCI gurus on this list, can somebody give suggestions as to why this card might not work with the default configuration? I feel like such an idiot for changing the BIOS settings knowing the troubled history of this configuration, but I was surprised that, after several hours of testing different things, we were never able to get it to work again! I will try down-grading the flash BIOS to the earlier, non-Plug-and-Play BIOS on Monday if it still doesn't work... Is this just a strange interaction between an old pre-Triton (and possibly not PCI-2.0 compliant) motherboard, and older PCI cards? One final note: The Buslogic card has only two jumpers, to set the memory location for the BIOS. With both jumpers off, the motherboard is supposed to set the location, but with the Micron motherboard, it merely prints "PCI Error" or worse (screen corruption or refuses to boot). The card claims to be PCI 2.0. -- Jake
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199702240611.WAA01315>