Re: workaround for BIOS / CHS stuff

[Patrick J. LoPresti]

K.G. <address@hidden> writes:

> On 02 Jul 2004 08:55:37 -0400
> "Patrick J. LoPresti" <address@hidden> wrote:
> 
> > The interface is stable as of 2.6.7.  The fields are named
> > /sys/firmware/edd/int13_dev80/{legacy_max_cylinder,legacy_max_head,legacy_sectors_per_track}.
> > This is a 100% reliable way to determine the correct geometry to
> > use.  A backport to the 2.4 kernels is in progress.  The only
> > problem is mapping between BIOS device numbers (dev80) and Linux
> > devices...
> 
> So what, is this a truly 100% reliable way to get the geometry or
> is there a mapping problem between BIOS and Linux devices ? I don't
> understand how this can be reliable if you're not sure about the
> device you're asking the geometry for ?

The geometry values are correct.  Figuring out which disk they apply
to can be tricky, so "100%" was an overstatement.

But not much of one.  Although it is "theoretically impossible" to do
this mapping reliably, it is similarly "theoretically impossible" to
locate the system boot device.  Yet people manage to install boot
loaders.

There are several other reasons this is not likely to matter in
practice...  But I will not go into them right now, because a) I am at
work and b) it makes no difference.  The point is that Parted should
not be dealing with this at all.

This is a case where Parted would do well to adopt the engineering
principle of "Somebody Else's Problem" (aka. "Not My Job"; see
http://www.crosbycpr.com/content/features/galleries/funny/large/NOTMYJOB.JPG).

Let Somebody Else figure out the geometry and add an interface to pass
that info to Parted.  I am willing to contribute patches if the
maintainers will indicate what sort of interface they would accept
(command line switches? environment variables? new Parted
directives?).

 - Pat