Workbench 1.3 on the new A3000s

                             by Ross Molden 



   ¿? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿?


   I sort of promised you a contribution during one of our discussions at
  W.O.C. but I always seem to have other things to do ... like playing with
  the A3000 I bought!

   In the course of this play, I came upon Kevin Donoghoe's question in
  Megadisc 29 about the possibility of running Kickstart1.3 on the new A3000
  with V2.05 ROMs.  The answer (courtesy of Nic Wilson) suggested that the
  DOS "CPU" command would do the trick.

   Well, I studied the DOS manual description of this command, and I may have
  missed something, but, although it allows you to load a ROM image into
  32-bit RAM for faster execution, I could find no mention of an option to
  load a different version of Kickstart so I assume Nic really meant Dave
  Haynie's "SETCPU" command.  (The latest version I could find is V1.60 on
  Fish 400) This is similar in function to the CPU command but does have a
  few extra options including the ability to load a new Kickstart from a file
  or from a Kickstart disk.

   The documentation for this program is aimed at the Nic Wilsons of this
  world but I understood enough of it to be encouraged to "give it a go".  I
  don't have a Workbench 1.3 partition on my hard drive so my aim was to load
  Kickstart 1.3 or 1.2 from disk and then boot from the appropriate Workbench
  on floppy disk.

   Dave Haynie warns of the possibility of problems with the code running on
  A3000s since "...just because you help design a machine doesn't necessarily
  mean you have one to code and test on at home."

   Unfortunately he was right!  My initial attempts, although starting with
  promise, invariably fell in a heap when it came to the crunch of actually
  rebooting with the new Kickstart code.

   The relevant command syntax to load an external Kickstart from disk is of
  the form:

         SETCPU KICKROM DFn:

   When run, this results in the old "Kickstart Hand" appearing on screen.
  Feed it a Kickstart disk and it reads the code into memory and then does a
  warm boot .... or it tries to.  Every time I tried it the system hung in a
  loop of unsuccessful boot attempts culminating in a red screen.

   Refusing to be beaten by a mere machine, I persevered, trying many
  inspired(?) variations until eventually stumbling upon a procedure that
  works.


   Here's what you do:

   .  On initial startup, interrupt the boot-sequence by holding down
      both mouse buttons.

   .  This puts you into the boot selection menu.

   .  Select the "Advanced Options" gadget at the bottom of the screen.

   .  You will now be in the Advanced Options Boot Menu.

   .  Toggle the "Startup-sequence Enable" gadget to read "Disable".

   .  Select the "USE" gadget.

   .  You should now be back in the first Boot Menu screen.

   .  Select the "WB_2.x" gadget.

   .  The machine should boot and leave you in a CLI window.

   .  Run your SETCPU command being careful to supply the full path
      name if you don't have the command in your c: directory.

          e.g. (In my case) SYSTEM/SETCPU KICKROM DF0:

   .  After a few moments the Kickstart Hand should appear on the
      screen.

   .  Insert a Kickstart disk in the specified drive (DF0: in my case).

   .  While the Kickstart is loading, the colours in the screen image will
      cycle.

   .  On completion, and after a short delay, the Workbench Hand should
      appear.

   .  Insert the appropriate Workbench disk.

   .  The normal Amigados CLI screen will be opened and boot progress
      will be reported as usual but eventually a software error requestor
      will appear.

   .  Cancel the requester.

   .  The machine will GURU (Oh no!) but, DON'T PANIC!

   .  Press the Left Mouse Button and you're away laughing!


   I suspect that the problem is related to just one of the commands in the
  V2.05 startup-sequence and it is probably not necessary to by-pass the
  sequence completely, but it was quicker than editing out each command in
  turn until the culprit was identified.  The procedure works for Kickstart
  versions 1.2 and 1.3 but version 1.1 still hangs for some reason.

   Anyway, that's how I got around the problem and if anyone else has had the
  same experience maybe it will be of some assistance.




                                          Ross Molden,
                                          5 Roberts Street,
                                          Birkenhead.
                                          S.A.  5015.


   ¿? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿*? 35 ¿?