DPES-31080 geometry (Was: NetBSD/CyberSCSI mk1)

[Jerome Lovy <Jerome.LOVY%teaser.fr@localhost>]

Matthias Scheler wrote:

>In article <yam6960.1811.138221888%teaser.fr@localhost>,
>Jerome Lovy <Jerome.LOVY%teaser.fr@localhost> writes:
>> Which "buggy emulex driver" are you talking about ?

>About "src/sys/arch/amiga/dev/sfas.c".

>> (I know that "cybscsi.device" is quite buggy, ...

>Not half as buggy as the NetBSD driver.
...

>> Besides, as I already stated, I have been able to successfully
>> use both cybscsi.device and DPES-31080 under AmigaDos...)

>Yes, my DPES-31080 always worked fine under AmigaOS, too. But under NetBSD
>I got a SCSI bus lookup after a few seconds.

Thanks for the information.
After some desapointing experiments trying to build a kernel
with ADE + pmake under AmigaOs, I eventually gave up today.
I rather bought a second IDE drive.  I will then first
install NetBSD onto that drive, and later try to do something
about Emulex device driver, if I can...

Still I've got a remaining question about DPES-31080. What the
hell is the true geometry of that drive ? All the preparing
utilities I tried gave me different answers.  In particular,
HDToolBox (upon cybscsi.device) finds only one head whereas
some IBM WWW page says there are four of them... (too bad, that
WWW page doesn't mention anything about the number of
tracks/cylinders). Could it be that a wrong RDB configuration
gets NetBSD emulex driver lost ? (I tried many different
geometries without any better result anyway...)

Regards,
    Jerome