hallo Benni,

> > Conclusion: I'm astonished that NetBSD does not accept this AHDI
> >             partition information, it looks perfectly valid to me.
> >             I'll check the source code later this day (it's really
> >             time to go to bed now ;-) ).
> 
> If it helps you, disklabel under NetBSD (booted from floppy...)
> tells me, that I have *one* partition (sd0d) with 1057616 sectors size...
> And 4.2BSD as partition ID...

that's exactly what I would expect. netbsd does not accept your ahdi
partition table, that means that as far as netbsd is concerned this
partition table does not exist: netbsd assumes that there is only
`garbage' on your disk. a disk containing only garbage is an
uninitialised disk.
the d partition is created internally by netbsd. it has the same size
as the c partition (the whole disk).


shortly after finishing the current disklabel implementation, i
discovered a serious design flaw, a potential memory leak. today
i send a fixed implementation to Leo; i hope this will make it in
netbsd 1.1. there will also be a tos tool, which does the same
checks as the kernel. i'll send you a copy of this tool when it's
finished. if the result is positive, the netbsd 1.1 kernel will
accept your ahdi partition table. if the result is negative, at
least we will know why the partition table is rejected (maybe
there is some bogus check, allthough i couldn't find one today).

i didn't check the current implementation, as i expect it to be
replaced. anyway, a temporary bugfix wouldn't be much help to
you; you wouldn't be able to apply a patch to the kernel source
and recompile.


waldi

PS sorry for missing capitals, malfunctioning shift key.