A compilation of replies...

 On Thu, 20 Oct 1994 01:25:43 -0400, Ken Hornstein <kenh@entropic.com> said:

>>>>I can mount /dev/fd0a as a msdos device and copy stuff from it (hence
>>>>the output from fips) but disklabel says:
>>>>
>>>>disklabel: ioctl DIOCGDINFO: Invalid argument
>>
>>>I seem to remember this error will occur, but the disklabel command actually
>>>succeeds.
>>
>>But not for me :-(

>Err, how do you know it doesn't work?  BTW, are you talking about the floppy
>drive or the hard drve?

 This is for the floppy drive and I make the dangerous assumption that
1. no label being output and 2. the above error message, means that
something is wrong.

>>It reads my MS-DOS disk fine (barring that it complains every time
>>that it cannot find the disk label), but comes back with some juddery
>>data for my NetBSD drive (is it really 24MB?) given that both are 1048
>>cyl, WD2540.

>Well, if the NetBSD drive is _all_ NetBSD, then it won't have a MS-DOS
>disklabel.

 Well I gathered that, but I'm still rather shocked that a NetBSD
program (fdisk) can't read one of it's own disks, or is it just a
literal port of the MS-DOS program?

 btw, the console dumps out the following for `disklabel fd0' when it
contains a MS-DOS disk:

fd0: read failed st0 40<abnrml> st1 1<no_am> st2 0 cyl 0 head 0 sec 1
blkno 0 skip 0 cylin 0 status 40
fd0: seek failed st0 80<invld> cyl 1


[...after output of fdisk on a NetBSD drive and a MS-DOS drive...]
>I noticed the BIOS geometries aren't the same between the two disks; should
>they be?  You can use fdisk to set those as well.

 The BIOS says they are both 1048 cyl, 16 heads, 63 sect/track, NetBSD
probes the same information on booting.  If fdisk reports something
different then that's it's problem.  [Someone more MS-DOS knowledgable
than I said that the figures looked correct for a MS-DOS drive so I
have been presuming that it's just not been working for a NetBSD
drive]

 On Thu, 20 Oct 94 09:39:05 -0400, "John F. Woods" <jfw@ksr.com> said:

>And I, for similar reasons, use the fraudulent geometry enforced by the
>Adaptec driver (32 sectors per track, 64 heads per track, for 1MB

 We tried using 522 cyl, 32 heads and 63 cyl/track (as many MS-DOS
programs report) but we ended up in the same mess.



 A further snippet of information, is that after NetBSD has trampled
over the partition table - thus losing my colleagues MS-DOS partition,
MS-DOS fdisk et al all see a single partition of 24MB which is exactly
what I read for my NetBSD drive where I let it take over the whole of
my disk.

 My gut feel is that NetBSD is ignoring the partition table when it
starts up and is stamping just it's own partition back.

Cheers,

	Ian

-------------------------------------------------------------------------------
      Ian Fitchet			I.D.Fitchet@fulcrum.co.uk
  Fujitsu Fulcrum Telecommunications ltd., Fordrough Lane, Birmingham, B9 5FL