On Thu, Jun 16, 2005 at 08:07:34PM +0100, David Brownlee wrote:
> >>I agree this is not the best way to go.. I am not an
> >>expert in this area so I can't tell why not force LBA48 for
> >>all drives that support it. I just assume that it would
> >>have been done already if it was best for everybody :-)
> >
> >The problem is that not all controllers support LBA48 (and in addition,
> >using LBA48 when it's not needed is slower). I know there are systems
> >using LBA48-capable drives on non-LBA48 controllers here, because we had to
> >fix the driver to not use LBA48 when not needed for those systems.
> 
> 	Another option would be to force LBA48 for that block only on
> 	drives that support it. If we were starting from scratch that
> 	would probably be the way to go, with a quirk the other way.

You mean LBA (not LBA48) I guess ? Well, all drives but seagate support
it (the seagate drives are broken, legacy LBA supports addressing sectors
up to 128GB *inclusive*). And the other way round, the problem is with
controllers, not drive. And the problem in this case would be much more
serious (corruption of sector 0, most probably).

> 
> 	Actually, it might make sense to make the current quirk for all
> 	seagate	drives...

It looks like not all large seagate drives have this problem, only the 200
and 250GB versions. There are several 160G versions that don't seem affected.

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la difference
--