22 Nov
1998
22 Nov
'98
6:09 a.m.
Tim Shoppa <SHOPPA(a)trailing-edge.com> writes:
That would explain why I couldn't boot it, yep.
Partially. You wouldn't be able to boot it on a MicroVAX by typing "B
MUA0" even if it were complete. The reason is that the standard tape-making
script writes the VAX-11 bootstraps on the tape, not the MicroVAX ones. The
two are completely different. The big VAXen with front-end processors,
microcode consoles and such can't boot from a tape by themselves. The
bootstraps that appear on standard BSD distributions are designed to be
loaded by manually typing in a little program from the console in hex and
manually transferring control to it. The hex codes for 4 such programs (for
different tape drives and controllers) appear in the installation docs.
They cannot be ported to MicroVAXen, however, because they use some
features that exist only on big VAXen.
MicroVAXen, however, have tape boot capabilities built right into their
ROMs. Much easier for the installer, needless to say. Also needless to say,
the protocol the ROM tape boot code uses is completely different from the
one the BSD developers have crafted for their very special purpose.
Therefore, a tape needs a completely different bootstrap in order to be
directly bootable on a MicroVAX. One was written for 4.3BSD-Tahoe, and it
appears in the distributed /usr/mdec. There are two problems, however.
First, the standard tape-making scripts don't put it on the tape. Second,
it only supports KA630. When KA650 support was added, everything else was
updated accordingly, but this one was apparently forgotten. In theory, the
code looks generic enough to run on KA650 out of the box, but in practice
it has a check for SID and refuses to run if it's not 08 (MicroVAX ii).
Right now I don't know enough VAX assembly language to remove this check or
extend to accept 0A (CVAX) as well.
In any event, I am now very happy to now have a
bootable copy of
4.3-Reno. (Installing as I type, AAMAF.)
Do you mean the one in /usr/home/msokolov/43reno.vax? How did you get it
to boot on a MicroVAX? Did you pull /usr/mdec/tmscpboot out of the tarball
and make a MicroVAX boot tape?
We all know now that Michael's on a crusade for
4.3-Tahoe, so would it be
completely unreasonable to build 4.3-Tahoe from sources under 4.3-Reno?
It's the most reasonable approach I can think of at the moment.
That's close to what I'm doing. There are two differences, though.
First, I'm using Ultrix as my cross-compilation base, not 4.3BSD-Reno. (I
would say there is less of a gap between 4.3BSD-Tahoe and Ultrix than
between Tahoe and Reno. The latter is really huge, it's a gap between True
UNIX(R) and a bloated and POSIXized fallen one.) Second, what I will be
building won't be plain Tahoe, it will be Quasijarus1, i.e., Tahoe plus
KA650 support and shadow passwords from Reno and other improvements from
both later CSRG code and my own brain. SCCS will be the #1 tool in the
process.
And what stands in the way of reading around the bad
blocks on Kirk
McKusick's 43tahoe_cci distribution? Even though I know that Rick
Copeland doesn't have all the fancy tape recovery equipment I have in my
lab, is there some fundamental problem preventing the use of "mt fsr"
commands to skip the bad block(s) and recover the rest of the "src" and
"usr" tree?
If you do this you will still miss something. OTOH, if you go to the
4.3tahoe directory on Kirk's 2nd CD-ROM, you won't miss anything, since all
of /usr and /usr/src is there. I can bet that the files on that CD-ROM
match the ones on the tape byte for byte.
Sincerely,
Michael Sokolov
Phone: 440-449-0299 (Home) 216-217-2579 (Cellular)
ARPA Internet SMTP mail: mxs46(a)k2.scl.cwru.edu
Received: (from major@localhost)
by minnie.cs.adfa.edu.au (8.9.1/8.9.1) id RAA14402
for pups-liszt; Sun, 22 Nov 1998 17:04:17 +1100 (EST)
(envelope-from owner-pups(a)minnie.cs.adfa.oz.au)
X-Authentication-Warning: minnie.cs.adfa.edu.au: major set sender to
owner-pups(a)minnie.cs.adfa.oz.au using -f