-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE]
On Behalf Of Jordi Guillaumes i Pons
Sent: 02 January 2013 22:37
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] DMC11 in next simh version... looks nice, and
question
about TOPS-10
Al 02/01/13 23:32, En/na Paul_Koning at Dell.com ha escrit:
Why those two modes? Real DMCs are symmetrical, I would have
expected a simulation to do that too.
Just to clarify it, I didn't wrote that software :). I don't want to get
anyone's
merits. And, for your question, I also wonder why did the author wrote it
that way. I guess since it is TCP based one of the partners must put
itself in
listen mode and the other one has to do the initial connect, hence the
need
to differentiate between a "primary" or caller and a "secondary" or
listener.
UDP does not have that problem, but of course you can't (easily) tunnel
UDP
using -for instance- ssh.
I am main author of the code, although the original idea came from
Hans-Ulrich (Ulli) H lscher who has also done most of the testing and
initiated its move into the official SIMH code, and Mark Pizzolato has done
an amazing job getting it into SIMH and closer to the standard of other
devices.
The reason for primary and secondary modes is as Jordi says, because it uses
a TCP connection. I tried to make it symmetrical but it was not very
reliable and had problems when one end was behind an ISP which blocked the
incoming TCP connections. The easiest way was to for one end to connect and
the other to listen. I think some new features have been added to SIMH which
might make this better, but I didn't have those available to me when I first
wrote the code quite some time ago, perhaps a later release could change to
use this.
You can indeed use the simulated DMC11 to tunnel DECnet over the internet
without using the bridge, my router, Multinet or a hardware router (of
course your emulated node has to be a router though). My HECnet node VAX780
(5.8) is running the DMC11 emulation 24x7 on a Raspberry Pi, it is
configured to connect to Ulli's emulation as and when it is available and so
puts his nodes on HECnet when available.
As for speed, Ulli did some tests but I don't recall the results, certainly
faster than the original hardware. There is also a setting that lets you
artificially slow down the speed so that you can get closer to the
experience you would have had with the real hardware.
Please note that we have only tested it significantly on VAX 780 emulations.
I don't know enough about PDP11 and PDP10 to really test there. I have been
struggling particularly with PDP10, trying to get TOPS-10 to use it, but I
really don't know how and my attempt to build a new monitor with DMR11
emulation (almost identical to DMC11) failed at the link stage, but I don't
know why.
Regards
Rob
On 01/02/2013 05:25 PM, Gregg Levine wrote:
(BTW the "buffered console" feature in 4.0 is also awesome,
specially to run headless simulators. Now if we could do an
unattended RSX boot it would be wonderful!).
Yes, an unattended RSX boot WOULD be nice. ;)
You could probably do it with "expect". I did some work in that area
several years ago. I remember I got pretty far with it. I should try
to dig it up.
It certainly sounds interesting.
However Dave for you to find it will require a shovel as it was deeply buried.
You have no idea. Never having run Windows, thus without the periodic
"wiping clean" of life that tends to happen when the OS shits the bed, I
have stuff in my primary home directory that dates back to the 1980s.
It is a BIG filesystem.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
On 01/02/2013 05:32 PM, Paul_Koning at Dell.com wrote:
The upcoming SIMH version (4.0) comes with a virtual DMC11. It has
been merged to the head branch in the git repository, so I have
been playing with it a little bit. The simh team has done an
aswesome work again, so kudos for Bob, Mark and all the others...
It seems the DMCx are supported in pdp10, pdp11 and the VAXen. The
basic instructions to make it work are in the DOC files. Just to
make the story short:
- DMC11s work in pairs, just like the real devices. - One of the
DMCs has to be put in PRIMARY mode, the other one in SECONDARY. The
default is SECONDARY.
Why those two modes? Real DMCs are symmetrical, I would have
expected a simulation to do that too.
I haven't seen the code, but I would guess that it's a matter of
non-DMC-related administrivia, like which end initiates the TCP connection.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
Al 02/01/13 23:32, En/na Paul_Koning at Dell.com ha escrit:
Why those two modes? Real DMCs are symmetrical, I would have expected a simulation to do that too.
Just to clarify it, I didn't wrote that software :). I don't want to get anyone's merits. And, for your question, I also wonder why did the author wrote it that way. I guess since it is TCP based one of the partners must put itself in listen mode and the other one has to do the initial connect, hence the need to differentiate between a "primary" or caller and a "secondary" or listener. UDP does not have that problem, but of course you can't (easily) tunnel UDP using -for instance- ssh.
On Jan 2, 2013, at 4:33 PM, Jordi Guillaumes i Pons wrote:
Hello,
The upcoming SIMH version (4.0) comes with a virtual DMC11. It has been merged to the head branch in the git repository, so I have been playing with it a little bit. The simh team has done an aswesome work again, so kudos for Bob, Mark and all the others... It seems the DMCx are supported in pdp10, pdp11 and the VAXen. The basic instructions to make it work are in the DOC files. Just to make the story short:
- DMC11s work in pairs, just like the real devices.
- One of the DMCs has to be put in PRIMARY mode, the other one in SECONDARY. The default is SECONDARY.
Why those two modes? Real DMCs are symmetrical, I would have expected a simulation to do that too.
paul
On Wed, Jan 2, 2013 at 4:47 PM, Dave McGuire <mcguire at neurotica.com> wrote:
On 01/02/2013 04:43 PM, Cory Smelosky wrote:
(BTW the "buffered console" feature in 4.0 is also awesome,
specially to run headless simulators. Now if we could do an
unattended RSX boot it would be wonderful!).
Yes, an unattended RSX boot WOULD be nice. ;)
You could probably do it with "expect". I did some work in that area
several years ago. I remember I got pretty far with it. I should try
to dig it up.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
Hello!
It certainly sounds interesting.
However Dave for you to find it will require a shovel as it was deeply buried.
-----
Gregg C Levine gregg.drwho8 at gmail.com
"This signature fought the Time Wars, time and again."
Al 02/01/13 22:43, En/na Cory Smelosky ha escrit:
How is the performance?
I just played a little bit :) But I guess its far better than a real DMC-11 ;)
On 01/02/2013 04:43 PM, Cory Smelosky wrote:
(BTW the "buffered console" feature in 4.0 is also awesome,
specially to run headless simulators. Now if we could do an
unattended RSX boot it would be wonderful!).
Yes, an unattended RSX boot WOULD be nice. ;)
You could probably do it with "expect". I did some work in that area
several years ago. I remember I got pretty far with it. I should try
to dig it up.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
On 2 Jan 2013, at 16:33, Jordi Guillaumes i Pons <jg at jordi.guillaumes.name> wrote:
Hello,
The upcoming SIMH version (4.0) comes with a virtual DMC11. It has been merged to the head branch in the git repository, so I have been playing with it a little bit. The simh team has done an aswesome work again, so kudos for Bob, Mark and all the others... It seems the DMCx are supported in pdp10, pdp11 and the VAXen. The basic instructions to make it work are in the DOC files. Just to make the story short:
- DMC11s work in pairs, just like the real devices.
- One of the DMCs has to be put in PRIMARY mode, the other one in SECONDARY. The default is SECONDARY.
- Each of the DMCs has to be attached to a local port.
- Each of the DMCs has to be peered with the remote port.
Let's say we have a VAX780 running at 10.0.0.1 and another one running at 10.0.0.2. To link both simulators via DMC11s we have to:
In 10.0.0.1 INI file:
simh> set dmc0 enable
simh> set dmc0 peer=10.0.0.2:22222
simh> att dmc0 11111
In 10.0.0.2:
simh> set dmc0 enable
simh> set dmc0 linemode=primary
simh> set dmc0 peer=10.0.0.1:11111
simh> att dmc0 22222
Then, in both systems, once VMS is running:
NCP>define line dmc-0 state on
NCP>define cir dmc-0 state on cost 5
NCP>set line dmc-0 all
NCP>set cir dmc-0 all
(The systems must be routers if they also use the ethernet circuit)
In a few seconds, the circuit will go up and the adjacency will come alive.
Thank you for this information.
So we have now an alternative way to link our HECnet nodes via internet. The problem is the link has to be defined in both machines (so it does not help road warriors, a VPN link is still required). It uses TCP links, but if the IP link goes down it is smart enough to reconnect when it comes up again. According to the comments in the code it seems there are plans to add UDP transport. It would be even nicer if it could emulate somehow the multinet tunnel protocol :) but.
How is the performance?
Now the question. It seems the DMC11 device is supported in the KS10 emulation. Could it be posible to use it to DECNETize (and HECNETize) TOPS-10 over simh? I have not been able to find the docs to install DECNET-10 in a 7.04 monitor... Any TOPS-10 hacker around? :)
Hello,
The upcoming SIMH version (4.0) comes with a virtual DMC11. It has been merged to the head branch in the git repository, so I have been playing with it a little bit. The simh team has done an aswesome work again, so kudos for Bob, Mark and all the others... It seems the DMCx are supported in pdp10, pdp11 and the VAXen. The basic instructions to make it work are in the DOC files. Just to make the story short:
- DMC11s work in pairs, just like the real devices.
- One of the DMCs has to be put in PRIMARY mode, the other one in SECONDARY. The default is SECONDARY.
- Each of the DMCs has to be attached to a local port.
- Each of the DMCs has to be peered with the remote port.
Let's say we have a VAX780 running at 10.0.0.1 and another one running at 10.0.0.2. To link both simulators via DMC11s we have to:
In 10.0.0.1 INI file:
simh> set dmc0 enable
simh> set dmc0 peer=10.0.0.2:22222
simh> att dmc0 11111
In 10.0.0.2:
simh> set dmc0 enable
simh> set dmc0 linemode=primary
simh> set dmc0 peer=10.0.0.1:11111
simh> att dmc0 22222
Then, in both systems, once VMS is running:
NCP>define line dmc-0 state on
NCP>define cir dmc-0 state on cost 5
NCP>set line dmc-0 all
NCP>set cir dmc-0 all
(The systems must be routers if they also use the ethernet circuit)
In a few seconds, the circuit will go up and the adjacency will come alive.
So we have now an alternative way to link our HECnet nodes via internet. The problem is the link has to be defined in both machines (so it does not help road warriors, a VPN link is still required). It uses TCP links, but if the IP link goes down it is smart enough to reconnect when it comes up again. According to the comments in the code it seems there are plans to add UDP transport. It would be even nicer if it could emulate somehow the multinet tunnel protocol :) but.
Now the question. It seems the DMC11 device is supported in the KS10 emulation. Could it be posible to use it to DECNETize (and HECNETize) TOPS-10 over simh? I have not been able to find the docs to install DECNET-10 in a 7.04 monitor... Any TOPS-10 hacker around? :)
(BTW the "buffered console" feature in 4.0 is also awesome, specially to run headless simulators. Now if we could do an unattended RSX boot it would be wonderful!).