Michael,

   Your patch to the keyboard delay worked!  Thanks for your help.

   Unfortunately, that's not the end of my problems.  I'm still having two
   more problems.  The first is intermittent and I was able to slip by
   it and the second is more obstinant.  
   
   First Problem:
   ==============

   I get through to the part of the installation script where it asks me if
   I want to format my hard disk:
   
>THIS IS YOUR LAST CHANCE!!!
>
>(answering yes will format your root partition on sd0)
>Are you SURE you want NetBSD installed on your hard drive? (yes/no) yes

   After I answer 'yes', I get various error messages.
   Sometimes I get an fddone panic:

>Initializing / (root) filesystem, and mounting...
>/dev/rsd0a:     23250 sectors in 62 cylinders of 15 tracks, 25 sectors
>        11.4MB in 4 cyl groups (16 c/g, 2.93MB/g, 704 i/g)
>super-block backups (for fsck -b #) at:
>panic: fddone
>Stopped at 0x81428 unlk a6
>db>

   Other times I get 'corrupted track' on fd0.  Notice the problem with
   track -1(!):

>Initializing / (root) filesystem, and mounting...
>/dev/rsd0a:     23250 sectors in 62 cylinders of 15 tracks, 25 sectors
>        11.4MB in 4 cyl groups (16 c/g, 2.93MB/g, 704 i/g)
>fd0: corrupted track (63) data.
>fd0: corrupted track (63) data.
>super-block backups (for fsck -b #) at:
>fd0: corrupted track (63) data.
>fd0: corrupted track (-1) data.
> 32, 6064, 12096, 18128,
>fd0: corrupted track (-1) data.
>newfs: ioctl (WDINFO): Invalid argument
>newfs: /dev/rsd0a: can't rewrite disk label
>fd0: corrupted track (-1) data.
>Bus error
>
>Creating a fstab...mkdir: /mnt/etc: Read-only file system
>//install: cannot create /mnt/etc/fstab: directory nonexistent
>#

   At this point the kernel is still up and I can even re-run the install 
   script.  When I do so, things work better.  I'll include the
   entire output because it may be relevant to the second problem, described
   below.

>Initializing / (root) filesystem, and mounting...
>/dev/rsd0a:     23250 sectors in 62 cylinders of 15 tracks, 25 sectors
>        11.4MB in 4 cyl groups (16 c/g, 2.93MB/g, 704 i/g)
>super-block backups (for fsck -b #) at:
> 32, 6064, 12096, 18128,
>/dev/sd0a on /mnt type ufs (local)
>
>Creating a fstab...
>Now lets setup your /usr file system
>(Once a valid input for drive and partition is seen
>it will be FORMATTED and inserted in the fstab.)
>choices: sd0
>which drive do you want /usr on?
>sd0
>You have selected sd0
>here is a list of partitions on sd0
>#        size   offset    fstype   [fsize bsize   cpg]
>  a:    23250   307875    4.2BSD     1024  8192    16   # (Cyl.  821 - 882)
>  d:     7875      750      ADOS                        # (Cyl.    2 - 22)
>  e:   149625     8625      ADOS                        # (Cyl.   23 - 421)
>  f:   149625   158250      ADOS                        # (Cyl.  422 - 820)
>  g:    94500   331125    4.2BSD     1024  8192     0   # (Cyl.  883 - 1134)
>which partition would you like to format and have
>mounted as /usr? (supply the letter): g
>
>Initializing /usr filesystem, and mounting...
>/dev/rsd0g:     94500 sectors in 252 cylinders of 15 tracks, 25 sectors
>        46.1MB in 16 cyl groups (16 c/g, 2.93MB/g, 704 i/g)
>super-block backups (for fsck -b #) at:
> 32, 6064, 12096, 18128, 24160, 30192, 36224, 42256,
> 48288, 543420, 60352, 66384, 72416, 78448, 84480, 90512,
>newfs: ioctl (WDINFO): Invalid argument
>newfs: /dev/rsd0g: can't rewrite disk label
>/dev/sd0g on /mnt/usr type ufs (local)
>
>Adding to fstab... done.


   Second Problem:
   ===============

   I cannot do a mount_ados succesfully.  I first thought the problem was that
   I used the ADOS 3.1 *directory caching* filing system for the partion that
   I have my NetBSD stuff on.  So I reformatted using normal FFS.  I still get
   following problem (this is the *last* partition (sd0f) on my disk):

># cd /mnt
># mkdir ados
># mount_ados /dev/sd0f /mnt/ados
>adosfs: aget: bad primary type blk 74813
>mount_ados: mount: Invalid argument
>#

   And my work partition (the *middle* partition (sd0e),
   uses directory caching):

># mkdir work
># mount_ados /dev/sd0e /mnt/work
>adosfs: aget: bad primary type blk 74813
>mount_ados: mount: Invalid argument
>#

   And my Workbench-only partition (the *first* partition (sd0d),
   also uses directory caching):

># mkdir wb31
># mount_ados /dev/sd0d /mnt/wb31
>adosfs: aget: bad primary type blk 3938
>mount_ados: mount: Invalid argument
>#

   I have zero bytes between my partitions, if that makes any difference.
   Notice the 'primary type blk' is the same (74813) for my last and middle
   partitions.

   For completeness, here's the figures from HDToolBox.  I have three volumes
   named WB31, Work and Apps.  This is a Maxtor, hence the partitions are named
   mxN:

           Par Start End   File   Res
           Name Cyl  Cyl  System  Blk
           ---- ---  --- -------- ---

      WB31 mx0    2   22 DirCache   0
      Work mx1   23  421 DirCache   0
      Apps mx2  422  820   FFS      0


   Any ideas???  Do I need to have some reserved blocks?

Chris Cobb
ccobb@mnemonics.com