Pavel Roskin <[EMAIL PROTECTED]> wrote:
> On Wed, 22 Sep 1999, Rogelio M. Serrnao wrote:
>
> > I have an IBM PS2 77s with a 2.88 Mb disk drive. I dd'd GRUB stage1 and
> > stage2 on a 1.44 Mb diskette. When I try to boot from that diskette, all I
> > get is "Read Error".This does not happen with lilo. I intend to use GRUB
> > to boot Mach4. I am trying to port Mach4 to the MCA bus.
>
> This is already the second message within few days that confirms that the
> BIOS values for floppies cannot be trusted.
I'm curious about this error. In the past, the iterative probe trying
to detect a read failure usually worked pretty well (though to work
correctly with CDROMs and LS-120's you want to alter the final fallback
a bit, but not throw away the concept!).
> The first time BIOS thought that the disk A: is 1.2M while it was a 1.44M
> drive and a 1.44M diskette.
> The result was that wrong sectors were read and stage2 just hung without
> error messages.
>
> This time BIOS apparently believes that all diskettes in 2.88M drives
> should also be 2.88M
> Stage1 tries to read sectors above 18 and gets an error.
The point is that, with the iterative probe, both the stage1 and stage2
should not believe the BIOS anyway.
Has this been changed since the old GRUB 0.5? It was there for a
reason.
> Possible solutions:
Please don't store the disk geometry. There is very likely a way to make
a reasonable probe that works, and in general, the geometry you'd need to
program into it may not be discoverable under an OS not using the BIOS.
--
Erich Stefan Boleyn \_ <[EMAIL PROTECTED]>
Mad but Happy Scientist \__ http://www.uruk.org/
Motto: "I'll live forever or die trying" ---------------------------
