[net.sources] UNIVS: Unix for J-11 or 11/84

rartym@cybavax.UUCP (Richard Artym) (03/26/86)

    Does there exist a Unix distribution for PDP's running the J-11 CPU chip?
Going by the `Differences' section in DEC's Architecture Handbook, the J-11
looks mainly like an 11/44, but with bits of 11/24, 34, 40 and 45 thrown in!

    We are installing a  J-11-based Nissho N1100 board as an upgrade to our
PDP-11/34A (yes, we're in the museum business); we can run the '34 system on
the new CPU without any changes whatsoever, but in this mode we can only
access an eighth of the on-board 2 Mbytes of memory (i.e. the maximum an 11/34
expects to see).  This is a minor upgrade in itself, since the J-11 is a bit
faster than the '34 on most things and, unlike our '34, has an FPU.  However,
we won't see any major improvement until we gain access to the other 7/8 of
that memory!  If you have a suitable distribution, or know of one, I would
be more than happy to hear from you!

    Incidentally, I believe that the "new" PDP-11/84 is based on the J-11, so
anything that runs on that will do us fine!  (Also, if there is sufficient
similarity between any of 11/23, 34, 40, 45, 60 or 70, and the J-11 or 11/84,
then I may be able to make do with just the new machine support file, conf/m.c,
and the sys/* files that need alteration, if any, since our V7 sources have
all those options built in.)

    If you have any queries, comments, answers or suggestions regarding this
or related topics, I'd appreciate this type of feedback too.  Thanks!

-- 
vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
> Rich Artym, Elec.Eng., Univ.College of Swansea, Swansea SA2 8PP, Wales, UK. <
>  UUCP : {UK}!ukc!reading!cybavax!rartym ; JANET : rartym@swxa/000060601800  <
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

henry@utzoo.UUCP (Henry Spencer) (04/01/86)

> Going by the `Differences' section in DEC's Architecture Handbook, the J-11
> looks mainly like an 11/44, but with bits of 11/24, 34, 40 and 45 thrown in!
> ... if there is sufficient similarity between any
> of 11/23, 34, 40, 45, 60 or 70, and the J-11 or 11/84, then I
> may be able to make do with just the new machine support file, conf/m.c,
> and the sys/* files that need alteration, if any, since our V7 sources have
> all those options built in.)

Long odds you will find that the thing acts pretty much like a 70 as far
as Unix is concerned.  To V7 Unix, with the exception of one or two fine
details of which registers are printed in the event of a memory-system
error, a 44 looks *exactly* like a 70 with Unibus disks.  The "Unibus
disks" part requires a couple of small changes to the bootstrapping code,
which ignores the Unibus map since Bell Labs's 70s had Massbus disks and
45s didn't have the map.  The other differences between 70s and 44s are
in things that Unix does not use; quite likely the same is true of the J11.

(Make sure that all your device drivers allocate and free the Unibus map
properly; some of the older drivers don't bother and hence won't work on
a 22-bit machine.  Note also that the hp driver thinks it can tell whether
the controller is Massbus or Unibus by whether the CPU is a 70 or not,
which is wrong...)

The machine this is being typed on, utzoo, is a 44 which thinks it is a 70.
Who am I to disabuse it?  It's served long and faithfully...
-- 
				Henry Spencer @ U of Toronto Zoology
				{allegra,ihnp4,decvax,pyramid}!utzoo!henry