On Tue, 11 Jul 2000, Andries Brouwer wrote:
> On Mon, Jul 10, 2000 at 10:44:06PM -0700, H. Peter Anvin wrote:
> > Andries Brouwer wrote:
> > >
> > > If a program needs the partition starting point, it can use the
> > > HDIO_GETGEO ioctl which returns a
> > >
> > > struct hd_geometry {
> > > unsigned char heads;
> > > unsigned char sectors;
> > > unsigned short cylinders;
> > > unsigned long start;
> > > };
> >
> > Ah yes, this is what I was looking for. This should be sufficient for
> > the time (at least until 2^32 sectors become an issue (2 TB disks).
>
> Unfortunately it is an issue already today.
> Several people complained to me about problems with a larger
> than 1 TB RAID, where signedness problems start occurring.
> And if 1 TB is used today then 2 TB is very close nearby.
>
> Probably start will have to become a long long, and we need
> an audit of the kernel source to see where things will fail.
> If it turns out that the change is too complicated we can
> get temporary relief by using a larger unit, like 4K blocks
> instead of 512-byte sectors.
Hi Andries,
Have you followed the "GLOBS" model to reduce the addressing size issue?
If I understand this, there will be logical sectors of '512-byte' that
will define a GLOB == N*(512-byte) as the physical sector. I think this
is best described as a 'multiwrite wrapper'. There are sector and
subsector mapping......
It burns a few brain cells to follow this concept but it has a derivative
in AV-Streams.
Cheers,
Andre Hedrick
The Linux ATA/IDE guy
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Jul 15 2000 - 21:00:12 EST