*BSD News Article 8076


Return to BSD News archive

Newsgroups: comp.unix.bsd
Path: sserve!manuel.anu.edu.au!munnari.oz.au!network.ucsd.edu!usc!elroy.jpl.nasa.gov!decwrl!deccrl!news.crl.dec.com!dbased.nuo.dec.com!nntpd.lkg.dec.com!funlwb.stl.dec.com!theroo.stl.dec.com!sampson
From: sampson@theroo.stl.dec.com (Phil Sampson)
Subject: [386BSD] IDE BIOS params and reboot loops
Message-ID: <1992Nov23.010923.23092@funlwb.enet.dec.com>
Lines: 68
Sender: sampson@redbck.stl.dec.com (Phil Sampson)
Nntp-Posting-Host: redbck.stl.dec.com
Reply-To: sampson@theroo.stl.dec.com
Organization: Digital Equipment Corporation
Date: Mon, 23 Nov 1992 01:09:23 GMT


To: cc465@cleveland.Freenet.Edu (Steven M Ratliff)
Cc: 
Reply-To: sampson@theroo.stl.dec.com
Subject: Re: [386BSD] Infinite reboot with partitioned IDE drive




Hi,

I had the same problem on a Taiwanese clone. My config was 20Mb RAM, 386 with
and ST1144 as wd0 and a 200Mb Rodime as 2nd drive (Rodime not used for BSD).

I found I could install the 386BSD filesystem on the ST3144 without a hitch if
I used the whole disk, but got a panic (it was fast but I think it said) that
the disk was unlabeled and had a bad magic number if I tried to install the
distribution in a DOS partition.

Another symptom booting with the fixit diskette and looking at the partition
(the utility escapes me) shoed only and 'a' partion was built in the dos
partition. (whenit works you get three bsd partitions)

I finally figured MY cure. When I had bought the clone, the dealer installed
the ST1144 drive parameters in NVRAM with the PHYSICAL cylinders and sectors
heads etc. As you know, IDE drives do an internal "mapping" of cyclinders and
sectors and you can have the BIOS talk to them in a number of configs.

The clue I found was doing a disklabel -r wd0 and looking at the number of
sectors and cylinders that INSTALL had found. It was different to my NVRAM!

I can only think it read from the IDE drive. There were also a lot of blank
fields in the disklabel display. (these were filled in after the problem
was fixed.) The install process didn't care that it never found an ST1144
in disktab either.

>From a Seagate bulletin board you can get information about drives and
recommended NVRAM BIOS setups so I procured the ST1144 and it showed that
the optimum BIOS setup was in fact the number of cylinders (1001 and sec 17)
that disklabel showed rather than the (999 and 36) that the dealer had entered.

Changing the BIOS, re-FDISKing and re-installing 386bsd worked fine plus gave
me a couple of extra Meg for my trouble. If you have this problem, read your
disk docs and try an alternative geometry in your BIOS table. It seems that
IDE drives have a number of possibilities.

A clue MAY be to see what disklable -r wd0 READ from the drive and THEN enter
those figures as you BIOS Values and reformat.

Hope this helps out someone.

-Phil

--

Phil Sampson					   sampson@theroo.stl.dec.com
Customer Support Centre				   VK2JNT@VK2RWI.NSW.AUS.OC 
Digital Equipment Corporation (Australia) Pty Ltd
Sydney Australia



--

Phil Sampson					   sampson@theroo.stl.dec.com
Customer Support Centre				   VK2JNT@VK2RWI.NSW.AUS.OC 
Digital Equipment Corporation (Australia) Pty Ltd
Sydney Australia