Re: Why check legacy partition table for non-legacy HDs?

[H. J. Lu]

On Sun, Sep 15, 2002 at 07:11:22AM +1000, Andrew Clausen wrote:
> On Sat, Sep 14, 2002 at 01:35:44PM -0700, H. J. Lu wrote:
> > I updated
> > 
> > https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=72686
> 
> Parted gets it's BIOS geometry from linux in the first place...
> unless it is inconsistent with the partition table, which it
> is in your case.
> 
> You haven't convinced me that these semantics are bad, or that
> parted doesn't match these semantics.
> 
> 
> Basically: BIOS geometry is not well defined.  I don't care if it
> sometimes doesn't match what the BIOS says - it is better to keep
> the partition table consistent.  (eg: while doing data recovery
> on another machine)

I think I know what happened. That IDE drive was connected to a 3ware
ATA IDE RAID controller, which reported the geometry as 9732/255/63.
When I moved it to an IDE channel on MB, BIOS reports 155114/16/63.
That is why Linux kernel reports:

...
hdc: 156355584 sectors (80054 MB) w/1819KiB Cache, CHS=155114/16/63, UDMA(33)
...
Partition check:
...
 hdc: [PTBL] [9732/255/63] hdc1



H.J.