max@mthvax.cs.miami.edu (Max Southall) (11/25/90)
I've just finished working on a CompuAdd 20 MHz 286 machine with IDE interface builtin that had terrible problems running SCO Xenix 286. It came with an Award BIOS installed, and would generate PANIC and DOUBLE PANIC in multiuser mode anywhere from a few seconds after boot to a couple of hours at the longest, running 1 to 4 users. After obtaining technical documentation for Chips & Technology's NEAT chipset (the one utilized in the CompuAdd 220) I determined that the problem was likely in the incomplete number of options shipped with the systems's extended CMOS setup program by Award, XSET. Not being able to obtain the details of Award's CMOS configuration, I modified the JLS bilingual BIOS we had earlier developed for similar problems on a 386 and wrote an extended CMOS setup program that allows configuration and display of virtually all the CS8221B registers. After installing this new version of the BIOS, and experimentation with register values, the system now works flawlessly at 20 MHz. The lesson? There are bus timing problems associated with using some systems' implementations of IDE and Chips & Tech CS8221B chipsets that need to be addressed by using non-stock configuration of bus timing parameters. But, they can be solved even if SCO doesn't get back to you (they didn't in this case) and both the box manufacturer and Award BIOS supplier are unable to help. In any case, thanks to Alan Wendt (the box's frustrated owner) and Alex Pournelle (iconoclast at large) for providing the impetus for this solution. - Max Southall - max@mamia.UUCP | mthvax!mamia!max Hardware/Software Designer
alex@grian.cps.altadena.ca.us (Alex Pournelle) (11/29/90)
max@mthvax.cs.miami.edu (Max Southall) writes: >I've just finished working on a CompuAdd 20 MHz 286 machine with IDE >interface builtin that had terrible problems running SCO Xenix 286. It came >with an Award BIOS installed, and would generate PANIC and DOUBLE PANIC >in multiuser mode anywhere from a few seconds after boot to a couple of >hours at the longest, running 1 to 4 users. >Not being able to obtain the details of Award's CMOS configuration, I >modified the JLS bilingual BIOS we had earlier developed for similar problems >on a 386 and wrote an extended CMOS setup program that allows configuration >and display of virtually all the CS8221B registers. After installing this >new version of the BIOS, and experimentation with register values, the >system now works flawlessly at 20 MHz. Also, did you get/try the SCO SLS 133 (which fixes some RLL drive problems, and is popularly rumoured to fix other, AT/IDE drive problems) on this system? You can get it from SCO's BBS, or I could copy the disk I got for you (E-mail-ing it? Well, erm, I'm not very good at that--USmail/UPS would be easier). Other than selling your work to Award for a very great amount of money :-{=} and congratulating yourself for a job well done (``just hacked the BIOS''--the old hacker spirit lives!), have you considered doing a free/share/commercial-ware product to let users at that there BIOS from (say) DOS or (gasp!) XENIX? Can I have a copy if you do? I'd love such a hunk-a-code for the many BIOSes (Award and AMI: R U lis'ning?) who don't include advanced setup. Boy, would I! The only `advanced' setup I've seen is shareware with a return address of Taipei, Taiwan, and my Chinese-English (& their programming) is very good. >In any case, thanks to Alan Wendt (the box's frustrated owner) and Alex >Pournelle (iconoclast at large) for providing the impetus for this >solution. You are quite welcome--good thing that "nn" has a huge bug, or I wouldn't have seen this thankyou. (Me--eager for praise? Yes, me eager for praise, UGH. ;-) ``Wile E. Coyote: Iconoclast.'' It has a ring to it...) Keep on computing, Alex -- Alex Pournelle, freelance thinker Also: Workman & Associates, Data recovery for PCs, Macs, others ...elroy!grian!alex; BIX: alex; voice: (818) 791-7979 fax: (818) 794-2297 bbs: 791-1013; 8N1 24/12/3