>
> Fine. Noisily complain so the user can select between the values from a
> potentially broken HD_GETGEO and a value determined by examination would
> be better than what we have now. The HD_GETGEO calls are not omnicient,
> or even 100% correctly implemented. When they are (if they ever are) I
> will stop complaining. But if the kernel returns incorrect data
> regularly, the external programs should have an easy way of determining
> and using the correct data.
>
Perhaps we could have HD_GETGEO return more than one value, values should
come from the disk's partition table, the BIOS, the disk controller, and
the low level driver. Lilo, and fdisk (and others) could then just ask
which, if any, to use.
Bryn
--
PGP key pass phrase forgotten, \ Overload -- core meltdown sequence
again :( | initiated.
/ This space is intentionally left
| blank, apart from this text ;-)
\____________________________________