disk geometry of an IDE drive not/wrong detected

disk geometry of an IDE drive not/wrong detected

Post by Karl-Heinz Herrman » Mon, 08 Dec 1997 04:00:00



Hallo!

I recently put together a "new" box from older parts.  Right now it's
a Weitek 486 EISA Board (unfortunately i have only ISA Components so far)
a  Multi I/O card with 2 serial,parallel,Gameport, Floppy and one IDE-Port.
Main IC on the board: Goldstar Prime 2 9239, Card Number GW2760 [PX]
I don't have the documentation for the IO-Card.

I connected one Western Digital Caviar 1210 IDE-Drive (210M).
This drive has three Jumper positions: Master , Slave and CS (Cable Select?).
To my knowledge it should not be jumpered at all as the only drive on one
IDE-Channel. Is this correct?

The error behavior:

I boot  Linux (2.0.31) from a bootfloppy (dd-copied kernel image).
If I DON'T declare the IDE-drive to the BIOS, the floppy boots, I get
a boot message with the correct CHS=989/12/35 and Linux runs just fine.

If I declare the drive in the Bios-setup (same settings as detected),
I can boot from floppy, but the kernel message at boottime shows a
CHS=0/0/0.  After that I can use fdisk, after I specify the disk
geometry per expert setup, but I can't use lilo to write the
bootsector. It allways complains: Device 0x0302: Got bad geometry
0/0/0

I saw the commandline option for lilo (-f) to specify a disk geometry
file but I couldn't find any information on the syntax of this file
(/etc/disktab as default).  I tried to boot DOS (MS-DOS and openDOS)
from floppy, and the Computer hangs (MS) or says Devide Error and
hangs just the same (openDOS).  The drive put in my primary Linux-Box
(usually scsi) boots just fine on an On-board EIDE-Controller of my
Gigabyte Board (detects Mode 1, if somebody can use this info).  I
even restored the MBR with Dos-fdisk and made the drive DOS-bootable
again, but had the same problems in the 486.

To me it looks like the Bios won't give the correct geometry to the
operating Systems. Could this be the case? Has anybody an Idea how to
get this thing workin properly?  Or at least something to try?

I would rather boot from the Harddrive and not have to use a floppy
all the time.

(I tried different jumper-settings on the drive, so the "no jumper"
setting works fine in the other Computer and tried all 8 Jumpers on
the I/O-Controller.  Originally they were all on, and I changed one at
a time. They switch the various I/O-Channels on/off and determine the
COM IO-addresses as far as I can tell)

Thank you very much,

Karl-Heinz
--
--------------------------------------------------------
Karl-Heinz Herrmann, Stubenlohstr. 25, D-91052 Erlangen

 
 
 

1. Kernel detects wrong hard drive geometry

I looked through dejanews, but could find anything related to this
particular subject:

I have a 486/33MHz system with a 540 MB Quantum ProDrive LPS540A. I'm
installing Slackware 3.0 from CD-ROM. The BIOS does *not* support LBD
for HD geometries, but it doesn't seem to complain about having >1024
cylinders. (It's an AMI 486DX ISA BIOS from 06/06/92)

On bootup (from original installation book disk), it goes through the
kernel (1.2.13) messages and prints out (approximately):

  hda: QUANTUM LPS540A, 515MB with 96kb cache, 1048/16/63, MaxMult=8

The hard disk's actual geometry is 1120/16/59 (540MB). In the BIOS I
have it set at 1024/16/63 (504MB) for the DOS partition.

Question: Why does the kernel detect the hard drive geometry at 1048
cylinders?

Thanks in advance. I would prefer a response via email,
  wizard_auz at iname.com

2. Apache + PHP + mod_gzip

3. 2.2.x kernel sees wrong IDE hard drive geometry

4. emacs or vi

5. 4.1.1 sees IBM 20Gb IDE disk geometry wrong.

6. B&W gypsy (sun4m, renamed to voyager) STOPped operating.

7. SCSI and IDE Drives - can't detect IDE drives

8. FTP from Browser vs. shell question

9. IDE disk geometry not recognised

10. Disk Geometry on Second IDE Disk

11. SCSI disk drive geometry limits hard disk choices?

12. ide disk, not detected (2.2.6 and 2.2.13) is there a fix ?