Return to BSD News archive
Received: by minnie.vk1xwt.ampr.org with NNTP id AA854 ; Mon, 08 Feb 93 13:00:33 EST Path: sserve!manuel.anu.edu.au!munnari.oz.au!spool.mu.edu!nigel.msen.com!math.fu-berlin.de!unidui!du9ds4!prang From: prang@du9ds4.fb9dv.uni-duisburg.de (Juergen Prang) Newsgroups: comp.unix.bsd Subject: Re: IDE Drive Translation Date: 6 Feb 93 23:05:13 GMT Organization: Universitaet Duisburg GH Lines: 61 Message-ID: <prang.729039913@du9ds4> References: <1993Feb5.162455.1030@mcs.anl.gov> NNTP-Posting-Host: du9ds4.fb9dv.uni-duisburg.de toonen@mcs.anl.gov (Brian R. Toonen) writes: >Could someone please send me some information (or a location where >information can be obtained) on IDE drive translation. I have a >Micropolis 2112A 1001Mb IDE drive with 2034 cylinders, 16 heads, and >63 sectors/track. I had a 200Mb DOS partition, a 200Mb partition >which was reserved for NT and 600Mb for 386BSD. >Normally, the DC-680 IDE adapter translates the drive into 1017 >cylinders with 32 heads. In order to get 386BSD to work, I had to >disable the translation option on the IDE adapter. This caused DOS to >report that only a 503Mb drive was installed. This was not a problem >until I attempted to install NT. > Apparently, NT did not like the 600Mb 386BSD partition. While >trying to format the second parition, it destroyed the 386BSD >partition and then the format program died with complaints of lacking >memory (I have 16Mb). I removed the 600Mb partition and tried to >install NT again; this time it was successful. I can now reinstall >386BSD, but anytime I have to reformat NTFS (likely to occur), I will >have to remove 386BSD first. >The solution would appear to be to switch the drive back to translated >mode, but this would cause BSD to fail (being my favorite of the three >OS's, that would not be the desired effect). Any suggestions or >comments other than to remove NT would be greatly appreciated. Did you fdisk your drive after you turned off the translation mode ? The partition table contains the same limits in cylinder (1024) and sector counts (63) that you will find in the BIOS (INT13) interface. Hence a change in translation (and, I assume, CMOS entry) will result in a totally different disk view to the BIOS that should also be setup in the partition table. I don't know anything about VMS+1 (WNT ;-), but if it uses BIOS and/or partition sector information to find its own partition limits, this may be the reason for your problems. But if your partition sector is setup correctly for the non-translated drive, it may be a bug in WNT. 386BSD won't be bothered by this, because the wd driver uses direct information from the drive about its structure together with partition and BSD-disklabel information to find its limits, if 386BSD is not on the 1st partition (but this is also the reason why 386BSD has problems on translated drives). BTW, I encountered a strange behavior of Windows 3.1's 32-bit disk access. I entered the correct values (i.e. 1053 cyl.) for my Conner 540MB IDE drive into the CMOS table, only to be sure 386BSD won't have any problems. But this is in discrepancy to the partition table entry for 386BSD, stating the last cylinder as 1023 for the reasons mentioned above. This causes 32-bit access to complain about an error although this discrepancy is not related to the DOS partition. Hence I changed the CMOS table entry to 1023 and now Windows' 32bit access works. I hope this will not interfere with the disklabel entries for 386BSD, which are set up to access the drive under 386BSD up to cylinder 1052 !? Any person out there who knows ? -- Juergen Prang | prang@du9ds4.fb9dv.uni-duisburg.de University of Duisburg |******************************************** Electrical Engineering | Logic is a systematic method of coming Dept. of Dataprocessing | to the wrong conclusion with confidence