On 2012-06-07 01:29, Rob Jarratt wrote:
As I have hacked my own copy around a bit to make it work on Windows, can
you tell me what the fix was as I may find it hard to locate your change?
The port variable in the struct BRIDGE should be unsigned... :-)
What changes have you made? Maybe something that should be incorporated in my sources?
Johnny
Thanks
Rob
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-
hecnet at Update.UU.SE] On Behalf Of Johnny Billquist
Sent: 06 June 2012 23:43
To: hecnet at Update.UU.SE
Subject: [HECnet] The bridge program...
Hi all. I found a stupid bug in the bridge program, which I fixed.
Improved the code documentation slightly while I was at it. New version at
http://www.update.uu.se/~bqt/hecnet as usual.
Johnny
I've been running it out here on an old U10 running base I do builds on.
It's sol9 9/04 with studio 11 and 9_Recommendeds fed to it. 1gb ram and
a 40gb disk. SIMH hasn't crashed yet. Idle seems to be functional also,
it's idling at about 17% cpu on a 2Mb cache 400Mhz UltraSPARC-IIi
running at 440Mhz. CPU is a v12 impl, Mask 9.1 in prtdiag -v.
It's been fluctuating between 91% usage and 17% for the last few days.
Only other oddity is that the machine has swift with a second HME and
scsi on it (tape). No probs thus far. I have dedicated the internal
happy meal for the TAP devs.
Al.
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On
Behalf Of Brian Hechinger
Sent: Thursday, 7 June 2012 6:53 AM
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] Eeeewwwww
On 6/4/2012 8:51 PM, Dave McGuire wrote:
On 06/04/2012 12:13 PM, Brian Hechinger wrote:
SIMH 3.9 somehow kernel paniced my solaris server!
Wow.
-brian
ps: it's older solaris, but stil!
Holy cow! I have NEVER seen that happen. (which is why I run
it!)
How the heck did you do that?
Built and ran SIMH 3.9
3.8-1 works just fine (it's what i'm running now) and 3.9 runs fine in
my sol11 VM.
-brian
On 06/06/2012 06:11 PM, Johnny Billquist wrote:
By the way, as a warning...
I seem to remember that DECnet support now have been dropped from Linux.
So it might not be in there anymore, if you look at recent versions.
Huh? Nope, works fine here, on a two-week-old Mint installation.
(Mint is Ubuntu with Canonical's bad decisions un-done)
I've also had less than stellar results from trying to talk from Linux
to RSX. So it might not work absolutely right under all conditions.
Developers mostly (it not only) had VMS machines to test against...
I've had similar results talking to RSTS/E DECnet. From the little
bits of old mailing list traffic I've seen, I'd guess they'd be happy to
have it work with other platforms' DECnet implementations, but finding
people with machines to test against is tough outside of this crowd.
I plan to contact the developers when I have a little time and offer
to do some more formalized testing against RSTS/E and RSX and get them
feedback, and possibly fix some of the issues.
This really needs to happen.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
Can you talk to sol::
-P
On 06/06/2012 06:11 PM, Johnny Billquist wrote:
By the way, as a warning...
I seem to remember that DECnet support now have been dropped from Linux.
So it might not be in there anymore, if you look at recent versions.
Huh? Nope, works fine here, on a two-week-old Mint installation.
(Mint is Ubuntu with Canonical's bad decisions un-done)
I've also had less than stellar results from trying to talk from Linux
to RSX. So it might not work absolutely right under all conditions.
Developers mostly (it not only) had VMS machines to test against...
I've had similar results talking to RSTS/E DECnet. From the little
bits of old mailing list traffic I've seen, I'd guess they'd be happy to
have it work with other platforms' DECnet implementations, but finding
people with machines to test against is tough outside of this crowd.
I plan to contact the developers when I have a little time and offer
to do some more formalized testing against RSTS/E and RSX and get them
feedback, and possibly fix some of the issues.
This really needs to happen.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
On 06/06/2012 06:07 PM, Rob Jarratt wrote:
I reckon it could be done with a single interface.
Will take a look at the specs, a raspberry pi DECnet router would be great!
I haven't followed this whole thread since I'm in a busy time right
now, but just a quick thought...if it can't easily be done with a single
physical interface, perhaps a virtual interface might do the trick?
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
On 06/06/2012 04:53 PM, Brian Hechinger wrote:
SIMH 3.9 somehow kernel paniced my solaris server!
Wow.
-brian
ps: it's older solaris, but stil!
Holy cow! I have NEVER seen that happen. (which is why I run it!)
How the heck did you do that?
Built and ran SIMH 3.9
3.8-1 works just fine (it's what i'm running now) and 3.9 runs fine in
my sol11 VM.
Did it have something to do with the networking perhaps, maybe a weird
interaction with the Crossbow subsystem? We already know that part of
that doesn't do things exactly the way we need them to, I wonder if
there are other issues.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
Hmm. Sure. But if you go the Linux routing route, then you instead get
into the problem that it don't have a point-to-point connection to
another Linux box on the Internet...
If anyone makes a Linux implemenation, I'll suggest that it would be
made to support the Cisco/GRE and Multinet/UDP formats, and you could
also have Johnny's bridge on the other side....
-P
"actual DECnet router" meant "actual DECnet router" :-) Until now I thought
this was something only available in a VMS router node.
VMS not required... MIM is an area router running RSX. :-)
But yeah, "DECnet router" really meant just that.
T10/T20 is routing-IV, if you want to burn kernel memory it can be
area router, but, on the KI10 we had batter use of the memory, like
talking IP....
It would be a fun exercise to turn the bridge into an actual DECnet router.
One day, when I have a lot of time I might even think about trying that.
Feel free. The specifications are out there, so it's definitely not an
impossible task. But I guess it will take some work. It would be very
nice if the bridge did turn into a WAN router...
Bliss to C or a blisse X86 code generator... ---:)
-P
On 2012-06-06 11:26, Peter Coghlan wrote:
No, you have not misunderstood. You are just not getting the full
picture. Yes, the Cisco box will transport the DECnet frames inside
something. At the bottom there will be IP, but most likely the DECnet
frames are actually sitting inside UDP...
My understanding of it is that GRE is another IP protocol on the same level
as TCP and UDP rather than using UDP.
You could definitely be right on GRE. It was way too long since I
actually tried working with GRE to really remember much anymore...
GRE, L2TPv3, TCP, UDP, ICMP are all protocols transported in/by
IP. It's a 8 bit identifier.
www.iana.org/assignments/protocol-numbers/protocol-numbers.xml
--P
(what is 120...)
As I have hacked my own copy around a bit to make it work on Windows, can
you tell me what the fix was as I may find it hard to locate your change?
Thanks
Rob
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-
hecnet at Update.UU.SE] On Behalf Of Johnny Billquist
Sent: 06 June 2012 23:43
To: hecnet at Update.UU.SE
Subject: [HECnet] The bridge program...
Hi all. I found a stupid bug in the bridge program, which I fixed.
Improved the code documentation slightly while I was at it. New version at
http://www.update.uu.se/~bqt/hecnet as usual.
Johnny
I've got a Windows NT 4.0 VM up occasionally too..
Sampsa
On 7 Jun 2012, at 00:00, Johnny Billquist wrote:
Just some fun details...
As of today, there are 321 nodes in the nodename database.
They are spread out over 16 areas.
We have machines located on (at least) three continents, if I remember right.
While not online all the time, I think we currently have atleast the following OSes represented:
RSX
RSTS/E
VMS
Ultrix
Linux
OSF/1
TOPS-10
Tops-20
Windows XP
IOS
If you know of any errors in this information, more fun facts, or anything else you'd like to share, feel free to do so.
Johnny
p.s. Retrochallenge registration is open. It runs for the month of July.
Would be great to see some entries from you guys.
http://retrochallenge.org
Regards, Mark
Hmmm... Maybe a DECnet router on a Raspberry Pi....
By the way: any more thought about where and when for DEC Legacy 2012?
Regards
Rob
Just some fun details...
As of today, there are 321 nodes in the nodename database.
They are spread out over 16 areas.
We have machines located on (at least) three continents, if I remember right.
While not online all the time, I think we currently have atleast the following OSes represented:
RSX
RSTS/E
VMS
Ultrix
Linux
OSF/1
TOPS-10
Tops-20
Windows XP
IOS
If you know of any errors in this information, more fun facts, or anything else you'd like to share, feel free to do so.
Johnny
Hi all. I found a stupid bug in the bridge program, which I fixed. Improved the code documentation slightly while I was at it. New version at http://www.update.uu.se/~bqt/hecnet as usual.
Johnny
On 05/06/12 23:04, Mark Benson wrote:
On 5 Jun 2012, at 19:54, Dave McGuire wrote:
UNIX "sucks" because you don't know how to use it, and its design
differs from the OS that you DO know how to use? Interesting logic. ;)
*poke poke*
Nooow Dave, that kind of humour has gotten you into all sorts of trouble before on ClassicCMP ;) ;)
Besides, anyone who thinks I know how to use VMS is a moron, Im a total VMS noob. I know properly 10x about Linux what I do about VMS. That said there are plenty areas I've never had to deal with and mounting without root privileges is seemingly one of them. Thing is it 'just works' in RSX and VMS it 'just don't work' in Linux, at least at a prompt using the conventional tools.
The specifics of how to accomplish this does, however, differ from
UNIX to UNIX. Under any modern-ish Linux system for example, this is
done automatically upon device insertion by a combination of udev, dbus,
and gvfs. It Just Works, I've done it five or six times since lunch
today. Of course you have to be logged into a "desktop" session in
order for it to work, but 99% of the time, that's what's going on.
Yes. Same as it works just fine in Mac OS X (except when Apple's crummy USB drivers cause another kernel panic). I appreciate that, but because my SimH emulations don't need a GUI... I don't need a GUI. To me if it doesn't work at the command prompt it doesn't work full stop. Stuff that happens in GUIs is written for GUIs, I don't consider GUIs that are often tailored to each distro part of the GNU Linux or UNIX, they are an extra. I'm old fashioned, so sue me.
If you want to do it a different way, say when nobody is logged in,
this is easily accomplished with a udev rule that matches the device
when it's inserted, and takes some action, in this case mounts the
filesystem. If you need to do that, let me know, and I may be able to help.
What I need to do is type 'mount<device> ~/usb' and have it work, without using sudo or su.
To my mind this is the equivalent of ALL DUA0: // MOU DUA0: MARKDISK1 MARK: in VMS... except you don't have top putz about in VMS, it just works.
That was my point, which in itself was kinda tongue-in-cheek anyway :)
Hey guys,
My email client has been slowly squirrelling HECnet emails away in a folder using a rule I set up and then forgot about since I got my shiny new iPad. Is there a cross platform way of achieving uniformity with this kind of thing using gmail hosted mail? Or do I need to configure each client. That'll be a PITA.
Anyway, I'm slowly catching up on the Raspberry Pi/SIMH thread. I had shipping notification yesterday, and this seems like an excellent use of the hardware. Glad to hear it's been an easy ride...
This is also mildly interesting: http://netbsd0.blogspot.co.uk/2012/06/retrocomputing-with-vamp-stack-vax.ht…
Always good to see a VAXstation being put to good use.
Now, more importantly, when do we start on that multi-GHz nVAX FPGA implementation?
Regards,
Mark
p.s. Retrochallenge registration is open. It runs for the month of July. Would be great to see some entries from you guys.
http://retrochallenge.org
Regards, Mark
--
--
Mark Wickens
http://wickensonline.co.ukhttp://declegacy.org.ukhttp://retrochallenge.orghttps://twitter.com/#!/@urbancamo
By the way, as a warning...
I seem to remember that DECnet support now have been dropped from Linux. So it might not be in there anymore, if you look at recent versions.
I've also had less than stellar results from trying to talk from Linux to RSX. So it might not work absolutely right under all conditions. Developers mostly (it not only) had VMS machines to test against...
Johnny
On 2012-06-07 00:07, Rob Jarratt wrote:
I reckon it could be done with a single interface.
Will take a look at the specs, a raspberry pi DECnet router would be great!
Regards
Rob
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-
hecnet at Update.UU.SE] On Behalf Of Mark Benson
Sent: 06 June 2012 22:41
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] Multinet Tunnel Connections to SG1::
On 6 Jun 2012, at 22:37,<Paul_Koning at Dell.com> wrote:
On Jun 6, 2012, at 5:34 PM, Johnny Billquist wrote:
On 2012-06-06 22:40, Rob Jarratt wrote:
...
It would be a fun exercise to turn the bridge into an actual DECnet
router.
One day, when I have a lot of time I might even think about trying
that.
Feel free. The specifications are out there, so it's definitely not an
impossible task. But I guess it will take some work. It would be very nice
if the
bridge did turn into a WAN router...
Johnny
DECnet Phase IV is pretty straightforward. The specs are all out there.
And
much of it is implemented in Linux, so you can find C sources ready to
go...
Even if sample code only does L1 routing, you're nearly there because area
routing is essentially the same thing done a second time on a second set
of
tables.
There may be mileage in building dedicated DECnet bridges/routers using
RaspberryPi boards in a suitable case running a very cut-down Linux with
an
advanced version of the bridge? I am no programming wizard but I am good
at compiling and testing stuff.
Would it require 2 Ethernet interfaces to work or would (like the bridge)
just
one work okay?
--
Mark Benson
http://DECtec.info
Twitter: @DECtecInfo
HECnet: STAR69::MARK
Online Resource& Mailing List for DEC Enthusiasts.
I reckon it could be done with a single interface.
Will take a look at the specs, a raspberry pi DECnet router would be great!
Regards
Rob
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-
hecnet at Update.UU.SE] On Behalf Of Mark Benson
Sent: 06 June 2012 22:41
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] Multinet Tunnel Connections to SG1::
On 6 Jun 2012, at 22:37, <Paul_Koning at Dell.com> wrote:
On Jun 6, 2012, at 5:34 PM, Johnny Billquist wrote:
On 2012-06-06 22:40, Rob Jarratt wrote:
...
It would be a fun exercise to turn the bridge into an actual DECnet
router.
One day, when I have a lot of time I might even think about trying
that.
Feel free. The specifications are out there, so it's definitely not an
impossible task. But I guess it will take some work. It would be very nice
if the
bridge did turn into a WAN router...
Johnny
DECnet Phase IV is pretty straightforward. The specs are all out there.
And
much of it is implemented in Linux, so you can find C sources ready to
go...
Even if sample code only does L1 routing, you're nearly there because area
routing is essentially the same thing done a second time on a second set
of
tables.
There may be mileage in building dedicated DECnet bridges/routers using
RaspberryPi boards in a suitable case running a very cut-down Linux with
an
advanced version of the bridge? I am no programming wizard but I am good
at compiling and testing stuff.
Would it require 2 Ethernet interfaces to work or would (like the bridge)
just
one work okay?
--
Mark Benson
http://DECtec.info
Twitter: @DECtecInfo
HECnet: STAR69::MARK
Online Resource & Mailing List for DEC Enthusiasts.
On 2012-06-06 23:40, Mark Benson wrote:
On 6 Jun 2012, at 22:37,<Paul_Koning at Dell.com> wrote:
On Jun 6, 2012, at 5:34 PM, Johnny Billquist wrote:
On 2012-06-06 22:40, Rob Jarratt wrote:
...
It would be a fun exercise to turn the bridge into an actual DECnet router.
One day, when I have a lot of time I might even think about trying that.
Feel free. The specifications are out there, so it's definitely not an impossible task. But I guess it will take some work. It would be very nice if the bridge did turn into a WAN router...
Johnny
DECnet Phase IV is pretty straightforward. The specs are all out there. And much of it is implemented in Linux, so you can find C sources ready to go... Even if sample code only does L1 routing, you're nearly there because area routing is essentially the same thing done a second time on a second set of tables.
There may be mileage in building dedicated DECnet bridges/routers using RaspberryPi boards in a suitable case running a very cut-down Linux with an advanced version of the bridge? I am no programming wizard but I am good at compiling and testing stuff.
Hmm. Sure. But if you go the Linux routing route, then you instead get into the problem that it don't have a point-to-point connection to another Linux box on the Internet...
Would it require 2 Ethernet interfaces to work or would (like the bridge) just one work okay?
It would (today) require two ethernet interfaces, in which it routes between. However, talking about connections across the world, you are not even sitting on any shared ethernet networks, so it will not give you what you need for HECnet until you add some point-to-point connection, using IP (and whatever) to some other destination.
Maybe something already exists to accomplish this. Not sure if the GRE implementation in Linux can be used to solve it...
Johnny
On 6 Jun 2012, at 22:37, <Paul_Koning at Dell.com> wrote:
On Jun 6, 2012, at 5:34 PM, Johnny Billquist wrote:
On 2012-06-06 22:40, Rob Jarratt wrote:
...
It would be a fun exercise to turn the bridge into an actual DECnet router.
One day, when I have a lot of time I might even think about trying that.
Feel free. The specifications are out there, so it's definitely not an impossible task. But I guess it will take some work. It would be very nice if the bridge did turn into a WAN router...
Johnny
DECnet Phase IV is pretty straightforward. The specs are all out there. And much of it is implemented in Linux, so you can find C sources ready to go... Even if sample code only does L1 routing, you're nearly there because area routing is essentially the same thing done a second time on a second set of tables.
There may be mileage in building dedicated DECnet bridges/routers using RaspberryPi boards in a suitable case running a very cut-down Linux with an advanced version of the bridge? I am no programming wizard but I am good at compiling and testing stuff.
Would it require 2 Ethernet interfaces to work or would (like the bridge) just one work okay?
--
Mark Benson
http://DECtec.info
Twitter: @DECtecInfo
HECnet: STAR69::MARK
Online Resource & Mailing List for DEC Enthusiasts.
On Jun 6, 2012, at 5:34 PM, Johnny Billquist wrote:
On 2012-06-06 22:40, Rob Jarratt wrote:
...
It would be a fun exercise to turn the bridge into an actual DECnet router.
One day, when I have a lot of time I might even think about trying that.
Feel free. The specifications are out there, so it's definitely not an impossible task. But I guess it will take some work. It would be very nice if the bridge did turn into a WAN router...
Johnny
DECnet Phase IV is pretty straightforward. The specs are all out there. And much of it is implemented in Linux, so you can find C sources ready to go... Even if sample code only does L1 routing, you're nearly there because area routing is essentially the same thing done a second time on a second set of tables.
paul
On 2012-06-06 22:40, Rob Jarratt wrote:
No, you have not misunderstood. You are just not getting the full picture.
Yes, the Cisco box will transport the DECnet frames inside something. At
the
bottom there will be IP, but most likely the DECnet frames are actually
sitting
inside UDP...
So what is the difference from the bridge program, you ask.
Well, simple, the bridge program tries to be clever about which frames to
send where, but it cannot be clever enough at times, and also there are
times when it is forced to send traffic that we might want to limit.
The reason is that the bridge is not really a part of the DECnet traffic,
but is
just relaying packets back and forth, making the two ends of the bridge
appear as if they were on the same ethernet segment.
The Cisco box, on the other hand, is a DECnet routing node. It sits on the
ethernet, and participate fully in the DECnet traffic.
Thanks. Now I get it. I didn't know that anyone other than DEC ever made
something that could route actual DECnet and therefore did not realise that
"actual DECnet router" meant "actual DECnet router" :-) Until now I thought
this was something only available in a VMS router node.
VMS not required... MIM is an area router running RSX. :-)
But yeah, "DECnet router" really meant just that.
It would be a fun exercise to turn the bridge into an actual DECnet router.
One day, when I have a lot of time I might even think about trying that.
Feel free. The specifications are out there, so it's definitely not an impossible task. But I guess it will take some work. It would be very nice if the bridge did turn into a WAN router...
Johnny
On 6/6/2012 5:26 AM, Peter Coghlan wrote:
No, you have not misunderstood. You are just not getting the full
picture. Yes, the Cisco box will transport the DECnet frames inside
something. At the bottom there will be IP, but most likely the DECnet
frames are actually sitting inside UDP...
My understanding of it is that GRE is another IP protocol on the same level
as TCP and UDP rather than using UDP.
You are correct. GRE is Layer 4 just like TCP and UDP.
-brian
On 6/4/2012 8:51 PM, Dave McGuire wrote:
On 06/04/2012 12:13 PM, Brian Hechinger wrote:
SIMH 3.9 somehow kernel paniced my solaris server!
Wow.
-brian
ps: it's older solaris, but stil!
Holy cow! I have NEVER seen that happen. (which is why I run it!)
How the heck did you do that?
Built and ran SIMH 3.9
3.8-1 works just fine (it's what i'm running now) and 3.9 runs fine in my sol11 VM.
-brian
On 6/6/2012 9:04 AM, Peter Lothberg wrote:
Could this work in a domestic environment with an ISP that
gives out dynamic IP addresses?
My head hurts.. Ask for a /56 of IPv6 space. -:)
If there are many users that need this, I will look in to it, I thnk
there was magic that could be applied.
On the side with the dhcp address it's simple. You just make the source the actual interface and not an IP.
interface Tunnel0
description Tunnel for DECnet to Dave M.
no ip address
decnet cost 10
tunnel source FastEthernet0/1
tunnel destination 50.73.179.1
tunnel path-mtu-discovery
end
interface FastEthernet0/1
description Internet Connection (Cable Modem)
ip address dhcp
ip nat outside
ip virtual-reassembly
duplex auto
speed auto
end
Now, in my case it's actually a static IP but RCN require me to actually fetch my IP via DHCP or it won't work.
As to how to handle this on the other end, I have no clue. :)
-brian
On 6/6/2012 8:48 AM, Peter Lothberg wrote:
On 2012-06-05 03:50, Peter Lothberg wrote:
For people who don't need LAT and have a Cisco box, this is way better
than the bridge program.
(If you really want LAT with other parts of HECnet, then there is no
alternative to the bridge...)
You can bridge LAT with the cisco box... -:) ((It might even HELP you
with LAT over WAN....)) (And it can act as a LAT termianl server)
Nice. I did not know that. That's even better. Will it just bridge the
protocols you ask for, or will it do them all? Hmm, I assume it's using
GRE for this?
Johnny
I -think- there where special stuff to help with LAT in WAN scenarios,
at some point there might even have been a telnet/lat gw.. I have to
go and refresh my memory..
SuperLAT! It no longer says that in the version banner though.
bart#lat ?
WORD Name of a remote system
bart(config-subif)#lat ?
enabled Enable LAT protocol translation
host Statically define LAT services
-brian
