17 Jul
2023
17 Jul
'23
12:56 p.m.
On Jul 17, 2023, at 12:28 PM, Henry Bent via cctalk
<cctalk(a)classiccmp.org> wrote:
Hi all,
I just noticed that images of a full RX50 floppy set for Ultrix-32m 1.2 was
posted on Bitsavers (
http://bitsavers.trailing-edge.com/bits/DEC/vax/ultrix/1.2/ULTRIX-32M_V1.2_…
). I am having difficulty parsing these images into a usable raw format
for SIMH.
As a reference, TUHS has a set of 1.0 floppies (
https://www.tuhs.org/Archive/Distributions/DEC/Ultrix-32M/ ) that are
usable for installation purposes. (You can ignore the 1.2 floppies in that
archive, they aren't actually a full set). The first disk of that
installer, 32m-1.0-bin/01, has a bootloader that starts at byte zero as we
would expect. This should be approximately equivalent to disk 1 in the
Bitsavers set. Oddly though, in the "raw" dump the bootloader doesn't
start until 0x1400, and a number of the other disks I looked at appear to
have odd holes/zeroes in them. IMD format dumps of the 1.2 disks are
provided but when I converted the IMD format to a raw image I got the same
issue.
It's unfortunate that there is no description of what format these image files have.
From your description it looks like these are physical track/sector ordered copies of the
floppy, while SIMH for the RX50 container files wants to see logical order. One tool that
can convert between the two is the "rx50.py" utility that is part of my RSTSFLX
V3 tool, a RSTS file system handler in Python. You can find it at
svn://akdesign.dyndns.org/flx/trunk (that's a Subversion repository). Note that the
rstsflx in the "simtools" repository of Open-SIMH is V2, that's an entirely
different version written in C. It also has the interleave algorithm but no command line
tool analogous to the rx50.py tool.
The key element is that logical track 0 is physical track 1, and similarly all the others
differ by one except logical track 79 which is physical track 0. That fits what you
stated: bootloader starting at 0x1400, that is sector 10 so the first sector in physical
track 1.
The other parts of the address translation are: sectors on the track are 2:1 interleaved,
and there is a 2 sector skew from each track to the next. So, for example, logical sector
0 is physical track 1 sector 0; logical sector 1 is physical track 1 sector 2; local
sector 5 is physical track 1 sector 1, and logical sector 10 is physical track 2 sector
2.
What you need to do is to invert that interleave algorithm to convert from the physical
sector order to the logical.
paul