On 14 Sep 2006 at 7:47, David Laight wrote:

> On Wed, Sep 13, 2006 at 10:38:08PM +0100, Peter Hardman wrote:
> > The cd doesn't have hexdump, and mounting the /usr slice and running from 
> > there, (or from hexdump on a floppy)  hexdump fails with a syntax error
> > - I guess it must need something else beside the executable.
> 
> Most particularly /libexec/ld.elf_so, chroot is the only way to find it.
> 
> > mbr_ext can't find the OS either...
> > 
> > I need to get hexdump working.  I'll have to put the IDE drive back
> > tomorrow and run hexdump from there.  I've not got into that level of detail
> > yet.  Its a nifty tool- you can read the all the gory details in the mbr (or
> > I could if my brain could dig out it's 386 machine code reference :)
> 
> In this case you only need to see if there is anything there at all.
> The 'gory details' or the mbr code are best looked at in its source.
> (And I've a long-standing offer of a beer for anyone who saves a byte
> from mbr_ext or mbr_bootsel while mainating the same functionality.)
Have you now. Something to do in those long winter evenings then :) Brings back 
memories of tuning my text code on the original W3.0 XGA driver to make it as 
fast as VGA hardware. 

I've run hexdump against the ATA drive and the SATA drive and sector 63 is 
identical on both drives. So I suppose we're back to the BIOS lying. 

Oh well. All I really wanted to do was to run bonnie++ to see if it was XP using the 
H/W badly or the H/W itself. 


> 

> 	David
> 
> -- 
> David Laight: david@l8s.co.uk
> 

Peter Hardman
Breeder of Shetland sheep and cattle