I was wrong. Call it a correctable memory error. Note I said "pretty good".Uh. You yourself said it was phase III before
3.0 is what shows up if you show exec cha. The NSP version doesn't match the DECnet phase in that case, which is what confused matters.
As I say, I had reason to dig into this more recently, and that resulted in better data. The bits say Phase II.
Here are the bits (These were captured by Rob): 58 01 19 06 54 4F 50 53 32 30 00 06 00 01 00 01 7F 00 03 00 00 03 00 00 00 This is not a phase III transport init. It's a Phase II NSP node init. >> 58 Startup message >> 01 Node init >> 19 Extensible binary node 25. >> 06 Image byte count >> 54 4F 50 53 32 30 T O P S 2 0 >> 00 No intercept functions >> 06 Requests: Intercept requested, no node verification >>00 01 Blocksize 256. >>00 01 NSPsize 256. >> 7F 00 Maxlinks 127. >> 03 00 00 Routing version 3.0.0 >>03 00 00 Comm version 3.0.0 >>00 SYSVER length (it's omitted by TOPS-20) Google AA-D600A-TC for the corresponding spec. This communication may not represent my employer's views, if any, on the matters discussed. On 02-Aug-13 13:20, Johnny Billquist wrote:
On 2013-08-02 19:08, Timothe Litt wrote:TOPS-20 for the KS is Phase II. Really. Besides having a pretty good memory, I have recently re-read the sources and have decoded the node initialization messages actually sent while working on the KDP.Uh. You yourself said it was phase III before. (http://mailman.trailing-edge.com/pipermail/simh/2013-April/007357.html)PMR is fine for getting from a Phase II node out past a Phase III. And yes, Phase II did support star topologies with so-called "intercept nodes" as routers (well, link forwarders.) Phase II messages have an optional routing header to support this.Yeah, PMR will do it. Still didn't make it "supported". :-)The KL did Phase III via the MCB (pdp-11 front ends), and eventually Phase IV over the KLNIA (ethernet). The phase III implementation actually had the MCB masquerade as the -20 (Phase III), while talking Phase II to the -20. Ugly, but at the time, the quickest way to get to Phase III. Phase IV is when the -20 itself actually became smart.Pretty much matches what I can recall, except that it was phase IV which was implemented in the FE, while the actual machine was still talking phase III. I guess I could look it up again, if it's really important.The TOPS-20 group abandoned the KS at Phase II, arguing that the KS couldn't support the newer, larger monitors. Marc Crispin managed to squeeze much of the newer code into a KS, but it was very low on address space and never released.MRC actually squeezed phase IV in. It sounds like you mixed up phase II and phase III here... KS was running phase III. As you said in the link I provided above, you yourself said as much before...I did not want TOPS-10 to fork, so I made TOPS-10 on the KS run DECnet Phase IV. TOPS-10 kept the KS going thru the end. (Well, thru the present.):-)Since I'm probably what's left of DEC in 36-bit land, I'm not too worried about 'official support' :-)Good point too.Yes, Phase III is a solid product, as is Phase IV. (But Phase V is an engineering marvel - that broke all the rules for a sensible user interface: complexity, compatibility. It's so unpopular that HP still offers Phase IV as an alternative on VMS today! It's a shame that it turned out that way.)Right. I know of plenty of people who "downgraded" back to phase IV.We'll see how far we can get with a VMS 3.something node - Rob thinks he can scare up a kit.Good luck! JohnnyThis communication may not represent my employer's views, if any, on the matters discussed. On 02-Aug-13 12:24, Johnny Billquist wrote:On 2013-08-02 13:02, Timothe Litt wrote:DEC shipped a product kit for DECnet that was simply a patch to enable it. It's kicking around the net.Are you talking about for VMS now, or something else...?Phase III is useful because Phase II nodes (and TOPS-20 on the KS is one) can talk to it. DECnet provided compatibility, but only between adjacent Phases...I'm not aware of anything talking much of phase II. I'm pretty sure T20 on a KS talks phase III, but I could be wrong. The documentation is out there, so it should be easy to verify. I know that DECnet-8 was only phase II, but I think everything else went beyond that.So with a Phase III node, one can have T20 -> PIII -> PIV -> PV -> anything.Sortof. Phase II nodes do not even understand routing, and expect all the nodes it want to talk to to be adjacent, not to mention the very limited number of nodes supported by phase II, and other oddities. (I think a topology like a star was supported, but I can't remember the details...) Also, it is still not supported to have a phase II node to to a phase IV node, even through an intermediate phase III node, if you want to talk about what DEC officially supported.But even just Phase III gets around the 2-sync line limitation of the -20.Right. Phase III is when things became sortof sane, even if there are still various limitations around.Either VMS or RSX would work for Rob's purposes. (I have the same requirement, but haven't tracked down an old VMS kit. ) So would TOPS-10 V7.02 - which I should have, but it's not on a disk I've restored as yet. Too many projects, not enough hours.:-) Time. Always a problem. JohnnyThis communication may not represent my employer's views, if any, on the matters discussed. On 02-Aug-13 06:28, Johnny Billquist wrote:On 2013-08-02 06:14, Cory Smelosky wrote:On Fri, 2 Aug 2013, Jarratt RMA wrote:I would like to get hold of a SIMH image of a Phase III DECnet node that wants to speak over a simulated KDP or DUP, both of which are currently being added to SIMH.I believe my best chance might be an RSX system, I have generated anRSX system once before but struggled a bit, so if anyone has something ready made it would be a big help. Thanks RobIf you have the license kit for 3.x, I could get Phase III going on VMS pretty quickly.Wouldn't phase III imply a system way before anything like a license manager facility existed? As for RSX, I doubt people in general would have kept such systems around just for the fun of it. Phase III is pretty old and long ago. Johnny _______________________________________________ Simh mailing list [email protected] http://mailman.trailing-edge.com/mailman/listinfo/simh
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ Simh mailing list [email protected] http://mailman.trailing-edge.com/mailman/listinfo/simh
