[Info-vax] MicroVAX 2000, SCSI, and OpenVMS 7.3

[Hans Vlems]

On 28 apr, 04:51, s... at obanion.us wrote:
> I know that many postings here relate to problems, so when I have a success, I post it for something different.
>
> In early 1993, I bought a used MicroVAX 2000 (KA410), with the intent of keeping my skills sharp when I quit my then current job (which I waited too long to do…), along with several pieces of it's specific documentation (Technical Manual, Maintenance Guide kit & Addendum, Customer Hardware).
>
> One of it's cost saving features was that it used the PC/XT era ST412/506 hard disk interface for MFM drives, although it had an early NCR 5380 SCSI chip that was only used as a tape interface to the TZK50 tape drive.  Also by the 1990s, SCSI had clearly taken over lower cost disks, including the lower cost VAXen, but no backward support in OpenVMS was available for the MV2000, although Trimarchi had advertised a SCSI disk for the MV2000 (I have a pamphlet somewhere…).
>
> If I understand correctly, by the late 1990s, Wolfgang J. Moeller had reverse engineered the SCSI port driver used on the MicroVAX3100, which was very similar, for use with the VM2000 including patching the boot ROM images, that he named PK2K.  The PK2K kit is still avail for download: Googling "pk2k vms" gives several hits, includingftp://ftp.gwdg.de/pub/vms/pk2k/
>
> The PK2K kit includes the source, objects, and images for VMS 5.5-2, 6.1, 6.2, 7.1 and 7.2, patch files for five VMS images used at boot time, and a patch file for use with the boot ROMs.  When HP recently began distributing VAX/VMS 7.3 for hobbyist, I found myself wondering if it would build in VMS 7.3.
>
> I should point out that I had burned and installed the patched boot ROMs some years ago, but never got around to updating the software.
>
> The driver source for 7.2 with the two lines of updates applied, built without error on a SIMH VAX running VMS 7.3, but the patch files for the five images in VMS failed.
>
> These five images are patched to support local booting (along with the patched ROMs) to replace the TVDRIVER and MUA support of the TZK50 for the PK2KDRVR and PKA port devices: CONFIGURE.EXE, STACONFIG.EXE, STADCONF.EXE, STASYSGEN.EXE, and SYSGEN.EXE.  The three STA images are used in building the standalone backup kit, where the other two are for normal booting, SYSGEN in particular.
>
> In the notes in the PK2K kit, DISM32 was used to disassemble the driver that the MV3100 used, and I figured that would be the place to start with these images.  With some luck, and recalling the pattern-based guesswork from the Crash Dump Analysis class I took at a OpenVMS bootcamp from Rob Eulenstein, I was able to find the new base addresses for the 5 patch files!  I don't really believe that the tables start at these addresses, but they do enable the relative offsets to work.
>
> I have yet to test the standalone backup kit, but SYSGEN and CONFIGURE are working correctly with two different SCSI disks (MAXTOR MXT 1240s and Quantum Fireball SE2.1S) and a TZ30.
>
> Below are the five patch file names and the one line for each that needs to be updated.
> Then the instructions in the PK2K kit can be followed.
>
> 2KCONFIGURE.COM
> define acf$devdesc      = 0CD17 ! old 0CC1D
>
> 2KSTACONFIG.COM
> define acf$devdesc      = 11A10 ! old 11810
>
> 2KSTANDCONF.COM
> define acf$devdesc      = 0D8E0 ! old 0D7E6
>
> 2KSTASYSGEN.COM
> define acf$devdesc      = 106F9 !  old 105FF
>
> 2KSYSGEN.COM
> define acf$devdesc      = 0E717 ! old 0E61D
>
> Notice that in all cases, the new base address is larger that the old.

That is an interesting post. I own two VAXstation 2000's. One houses a
dead RD54 and the other an RD53 and the latter still worked when it
was last booted, about a year ago. I'm aware of the work of Wolfgang
Moeller and what has kept me from applying the results is that you
must burn your own ROMs, right?
Is that a costly process? I'm all in favour of learning new things but
within the limits my wallet imposes...
Hans