From robert.jarratt at ntlworld.com Sat Nov 3 18:15:51 2012 From: robert.jarratt at ntlworld.com (Rob Jarratt) Date: Sat, 3 Nov 2012 22:15:51 -0000 Subject: [Simh] Slightly OT: Headless SIMH on Raspberry Pi Message-ID: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> I want to run SIMH on my headless Raspberry Pi, to do so I use putty to connect to the pi and log in. From the shell I can run SIMH just fine. What I would like though is for my session not to disappear if I close putty and to be able to reconnect to the session at some later time, with SIMH still running. I am not really a Unix expert, is there a way to do this? Regards Rob -------------- next part -------------- An HTML attachment was scrubbed... URL: From abs at absd.org Sat Nov 3 18:41:30 2012 From: abs at absd.org (David Brownlee) Date: Sat, 3 Nov 2012 22:41:30 +0000 Subject: [Simh] Slightly OT: Headless SIMH on Raspberry Pi In-Reply-To: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> References: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> Message-ID: On 3 November 2012 22:15, Rob Jarratt wrote: > I want to run SIMH on my headless Raspberry Pi, to do so I use putty to > connect to the pi and log in. From the shell I can run SIMH just fine. What > I would like though is for my session not to disappear if I close putty and > to be able to reconnect to the session at some later time, with SIMH still > running. I am not really a Unix expert, is there a way to do this? "screen" http://www.gnu.org/software/screen/ or "tmux" http://tmux.sourceforge.net/ are two very useful tools for just this :) From md.benson at gmail.com Sat Nov 3 19:18:28 2012 From: md.benson at gmail.com (Mark Benson) Date: Sat, 3 Nov 2012 23:18:28 +0000 Subject: [Simh] Slightly OT: Headless SIMH on Raspberry Pi In-Reply-To: References: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> Message-ID: On 3 Nov 2012, at 22:41, David Brownlee wrote: > On 3 November 2012 22:15, Rob Jarratt wrote: >> I want to run SIMH on my headless Raspberry Pi, to do so I use putty to >> connect to the pi and log in. From the shell I can run SIMH just fine. What >> I would like though is for my session not to disappear if I close putty and >> to be able to reconnect to the session at some later time, with SIMH still >> running. I am not really a Unix expert, is there a way to do this? > > "screen" http://www.gnu.org/software/screen/ or "tmux" > http://tmux.sourceforge.net/ are two very useful tools for just this > :) I use gnu 'screen' extensively on my Rasbperry Pi VAX setups to do just that. Works a charm, and you can detach the session at any time and it will keep buffering. Right now I recommend this as the best way to run SimH headless and without a need for a perpetual terminal. I'm hoping something better might come along that doesn't require additional software, though. -- Mark Benson http://DECtec.info Twitter: @DECtecInfo HECnet: STAR69::MARK Online Resource & Mailing List for DEC Enthusiasts. From robert.jarratt at ntlworld.com Sun Nov 4 02:27:40 2012 From: robert.jarratt at ntlworld.com (Rob Jarratt) Date: Sun, 4 Nov 2012 07:27:40 -0000 Subject: [Simh] Slightly OT: Headless SIMH on Raspberry Pi In-Reply-To: References: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> Message-ID: <028101cdba5d$e1b7b3d0$a5271b70$@ntlworld.com> > -----Original Message----- > From: simh-bounces at trailing-edge.com [mailto:simh-bounces at trailing- > edge.com] On Behalf Of Mark Benson > Sent: 03 November 2012 23:18 > To: simh at trailing-edge.com > Subject: Re: [Simh] Slightly OT: Headless SIMH on Raspberry Pi > > > On 3 Nov 2012, at 22:41, David Brownlee wrote: > > > On 3 November 2012 22:15, Rob Jarratt > wrote: > >> I want to run SIMH on my headless Raspberry Pi, to do so I use putty > >> to connect to the pi and log in. From the shell I can run SIMH just > >> fine. What I would like though is for my session not to disappear if > >> I close putty and to be able to reconnect to the session at some > >> later time, with SIMH still running. I am not really a Unix expert, is there > a way to do this? > > > > "screen" http://www.gnu.org/software/screen/ or "tmux" > > http://tmux.sourceforge.net/ are two very useful tools for just this > > :) > > I use gnu 'screen' extensively on my Rasbperry Pi VAX setups to do just that. > Works a charm, and you can detach the session at any time and it will keep > buffering. > > Right now I recommend this as the best way to run SimH headless and > without a need for a perpetual terminal. I'm hoping something better might > come along that doesn't require additional software, though. > > -- > > Mark Benson > > http://DECtec.info > Twitter: @DECtecInfo > HECnet: STAR69::MARK > > Online Resource & Mailing List for DEC Enthusiasts. > > _______________________________________________ > Simh mailing list > Simh at trailing-edge.com > http://mailman.trailing-edge.com/mailman/listinfo/simh Thanks to everyone for the replies. I am trying out screen now and it seems to be working well. Regards Rob From paulhardy2 at btinternet.com Sun Nov 4 13:50:28 2012 From: paulhardy2 at btinternet.com (Paul Hardy 2) Date: Sun, 4 Nov 2012 18:50:28 -0000 Subject: [Simh] Slightly OT: Headless SIMH on Raspberry Pi In-Reply-To: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> References: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> Message-ID: <000301cdbabd$422aaa10$c67ffe30$@btinternet.com> I use VNC to connect to my headless Raspberry Pi to run a SimH VAX- see (http://gettingstartedwithraspberrypi.tumblr.com/post/24142374137/setting-up -a-vnc-server). This survives disconnection OK. Paul. From: simh-bounces at trailing-edge.com [mailto:simh-bounces at trailing-edge.com] On Behalf Of Rob Jarratt Sent: 03 November 2012 22:16 To: simh at trailing-edge.com Subject: [Simh] Slightly OT: Headless SIMH on Raspberry Pi I want to run SIMH on my headless Raspberry Pi, to do so I use putty to connect to the pi and log in. From the shell I can run SIMH just fine. What I would like though is for my session not to disappear if I close putty and to be able to reconnect to the session at some later time, with SIMH still running. I am not really a Unix expert, is there a way to do this? Regards Rob -------------- next part -------------- An HTML attachment was scrubbed... URL: From oboguev at yahoo.com Sat Nov 10 15:38:55 2012 From: oboguev at yahoo.com (Sergey Oboguev) Date: Sat, 10 Nov 2012 12:38:55 -0800 (PST) Subject: [Simh] Multiprocessor VAX simulator In-Reply-To: <4CD39E08-92F1-44F5-A56C-21C92E2808A3@gmail.com> References: <4CD39E08-92F1-44F5-A56C-21C92E2808A3@gmail.com> Message-ID: <1352579935.10136.YahooMailRC@web184302.mail.ne1.yahoo.com> VAX MP is a derivative variant of SIMH VAX simulator capable to execute OpenVMS (VAX/VMS) in true SMP mode on multiprocessor host machines, including modern commodity multicore/hyperthreaded PCs and Macs. Initial release of VAX MP supports Intel x86 and x64 processors based machines as host hardware platform with Windows, Linux and OS X as host operating systems. The number of VAX MP virtual processors is limited to the number of host machine's simultaneous execution units (logical CPUs). For example, PC with quad-core i7 processor has hyperthreaded cores and would allow to run VAX MP instance with up to 8 virtual processors. VAX MP had been tested to OpenVMS VAX theoretical limit of 32 processors. VAX MP is an architectural simulator. The focus is on OpenVMS SMP capability, rather than simulation of specific historical hardware. Accordingly VAX MP does not attempt to simulate any historically existing VAX multiprocessor model, although it may be used as a stepping stone towards further development of such a simulator. Instead VAX MP simulates a multiprocessor variant of MicroVAX 3900, a machine that never historically existed, but complies generic VAX architecture and runs OpenVMS in SMP mode. VAX MP is currently downloadable from my pesonal web page that hosts source code, documentation, prebuilt binaries and screenshots. http://www.oboguev.net/vax_mp Those nostalgic for "real iron" may also view photographs of a session using VT320 connected via DECserver 200 to a VAXcluster of two VAX MP nodes, each with 8 processors, under heavy cluster IO traffic. Important legal notice: Although on a purely technical level VAX MP does not require VMS SMP extension license to run and OpenVMS runs fully functional in SMP mode on VAX MP without prompting for SMP extension license, however from the legal standpoint running VMS in SMP mode may require a license for VMS SMP extension. Whether such license is needed for hobbyist use is a gray area. Please be sure to consult "Legal information notice B" in "VAX MP OpenVMS User Manual" for more information regarding licensing situation. I personally do have SMP extension license legally sufficient to cover VAX MP development and testing "in the white", but your mileage may vary. I am unable to advise whether hobbyist use of OpenVMS VAX in multiprocessor mode would require SMP extension license. From oboguev at yahoo.com Sun Nov 11 16:09:17 2012 From: oboguev at yahoo.com (Sergey Oboguev) Date: Sun, 11 Nov 2012 13:09:17 -0800 (PST) Subject: [Simh] Virtual VAXen In-Reply-To: <1341840220.61205.YahooMailRC@web181119.mail.ne1.yahoo.com> References: <4FF9F720.2010207@supnik.org> <1341840220.61205.YahooMailRC@web181119.mail.ne1.yahoo.com> Message-ID: <1352668157.43838.YahooMailRC@web184305.mail.ne1.yahoo.com> I came across another article regarding the history of VAX virtualization project at DEC. Surprisingly, the projects dates back to 1981 and ran till 1990. The goal was to achieve A1 security level certification. Steve Lipner et al., "Lessons from VAX/SVS for High Assurance VM Systems" http://dx.doi.org/10.1109/MSP.2012.87 http://www.cse.psu.edu/~tjaeger/cse543-f12/docs/SP_SPSI-2012-03-0042 R1_Zurko.pdf ________________________________ From: Sergey Oboguev To: simh at trailing-edge.com Sent: Mon, July 9, 2012 6:23:41 AM Subject: Re: [Simh] Virtual VAXen > Ah yes, the VVAX option... another attempted graft on the VAX architecture >tree. Unlike vectors, it never got as far as a real implementation. According to VVAX / VAX Security Kernel developers, while it was never released externally as a product, there was internal proof-of-concept implementation on 11/730 and production quality implementation on 8800, the latter field-tested in 1989 at a number of external customers' sites. http://dx.doi.org/10.1109/32.106971 http://www.cse.psu.edu/~tjaeger/cse543-f06/papers/vax_vmm.pdf http://dx.doi.org/10.1109/RISP.1990.63834 http://www.scs.stanford.edu/nyu/02sp/sched/vmm.pdf And here is somewhat more detailed technical description: http://dx.doi.org/10.1109/ISCA.1991.1021630 http://mprc.pku.cn/mentors/training/ISCAreading/1991/p380-hall/p380-hall.pdf -------------- next part -------------- An HTML attachment was scrubbed... URL: From beebe at math.utah.edu Sat Nov 17 11:15:47 2012 From: beebe at math.utah.edu (Nelson H. F. Beebe) Date: Sat, 17 Nov 2012 09:15:47 -0700 (MST) Subject: [Simh] new papers on von Neumann's programs and machine Message-ID: [This posting is slightly off-target, because it is about virtual machines, and not Simh specifically. Nevertheless, the poster feels that it may well be of interest to list readers.] The latest issue of the IEEE journal Computer arrived in my mailbox a couple of days ago. It contains an interesting article that describes an emulator for John von Neumann's IAS (Institute for Advanced Study) computer, designed in the late 1940s, but not fully operational until 1952 (see, e.g., http://en.wikipedia.org/wiki/IAS_machine ). The article also describes the testing of several early programs written by von Neumann and his close collaborator, Herman Goldstine. The article authors report finding several bugs in the programs (all of which were published about five years before actual hardware became operational). Some of the errors arise from transcription from technical reports into the volumes of the Collected Works of John von Neumann but others were real bugs in the original programs. This demonstrates once again that unless published program code has actually been run on a digital computer, it is probably wrong. The new article contains further references, one of which is about the emulator itself, and is included below. There is also an important earlier third paper that is not cited by the new article. Here is a summary of those references, with links: Donald E. Knuth [John] von Neumann's first computer program ACM Computing Surveys 2(4) 247--260 December 1970 http://doi.acm.org/10.1145/356580.356581 Barry S. Fagin and Dale J. Skrien IASSim: a Programmable Emulator for the Princeton IAS/von Neumann Machine Proceedings of the 42nd ACM Technical Symposium on Computer Science Education (SIGCSE 11), 359--364 (2011) http://dx.doi.org/10.1145/1953163.1953271 http://www.cs.colby.edu/djskrien/CPUSim/ http://www.cs.colby.edu/djskrien/IASSim/ Barry Fagin and Dale Skrien Debugging on the Shoulders of Giants: von Neumann's Programs 65 Years Later [IEEE] Computer 45(11) 59--68 November 2012 http://doi.ieeecomputersociety.org/10.1109/MC.2012.69 ---------------------------------------- P.S. I maintain many bibliographies, one of which is on virtual machines: http://www.math.utah.edu/pub/tex/bib/virtual-machines.bib http://www.math.utah.edu/pub/tex/bib/virtual-machines.html [They look similar on the screen, but the HTML version has live hypertext links, and the .bib version is needed for BibTeX processing.] That bibliography was first created on 10-Apr-2006, and now has almost 1000 entries. Its contents are mined from an SQL database representing a collection of almost 700,000 entries covering several hundred journals, plus many author- and subject-specific bibliographies. There is also a complete bibliography of von Neumann's works in the BibNet Project archives: http://www.math.utah.edu/pub/bibnet/authors/v/von-neumann-john.bib http://www.math.utah.edu/pub/bibnet/authors/v/von-neumann-john.html ------------------------------------------------------------------------------- - Nelson H. F. Beebe Tel: +1 801 581 5254 - - University of Utah FAX: +1 801 581 4148 - - Department of Mathematics, 110 LCB Internet e-mail: beebe at math.utah.edu - - 155 S 1400 E RM 233 beebe at acm.org beebe at computer.org - - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe/ - ------------------------------------------------------------------------------- From lbickley at bickleywest.com Sat Nov 17 13:51:54 2012 From: lbickley at bickleywest.com (Lyle Bickley) Date: Sat, 17 Nov 2012 10:51:54 -0800 Subject: [Simh] new papers on von Neumann's programs and machine In-Reply-To: References: Message-ID: <20121117105154.2c779c49@asrock.bcwi.net> On Sat, 17 Nov 2012 09:15:47 -0700 (MST) "Nelson H. F. Beebe" wrote: > [This posting is slightly off-target, because it is about virtual > machines, and not Simh specifically. Nevertheless, the poster feels > that it may well be of interest to list readers.] > > The latest issue of the IEEE journal Computer arrived in my mailbox a > couple of days ago. It contains an interesting article that describes > an emulator for John von Neumann's IAS (Institute for Advanced Study) > computer, designed in the late 1940s, but not fully operational until > 1952 (see, e.g., > > http://en.wikipedia.org/wiki/IAS_machine > > ). > > The article also describes the testing of several early programs > written by von Neumann and his close collaborator, Herman Goldstine. > The article authors report finding several bugs in the programs (all > of which were published about five years before actual hardware became > operational). Some of the errors arise from transcription from > technical reports into the volumes of the Collected Works of John von > Neumann but others were real bugs in the original programs. This > demonstrates once again that unless published program code has > actually been run on a digital computer, it is probably wrong. > > The new article contains further references, one of which is about the > emulator itself, and is included below. There is also an important > earlier third paper that is not cited by the new article. Here is a > summary of those references, with links: > > Donald E. Knuth > [John] von Neumann's first computer program > ACM Computing Surveys 2(4) 247--260 December 1970 > http://doi.acm.org/10.1145/356580.356581 > > Barry S. Fagin and Dale J. Skrien > IASSim: a Programmable Emulator for the Princeton IAS/von > Neumann Machine Proceedings of the 42nd ACM Technical Symposium on > Computer Science Education (SIGCSE 11), 359--364 (2011) > http://dx.doi.org/10.1145/1953163.1953271 > http://www.cs.colby.edu/djskrien/CPUSim/ > http://www.cs.colby.edu/djskrien/IASSim/ Thanks for sharing these links! I've been playing with the IAS Simulator, and it is an excellent tool to become familiar with what it was like to program the IAS! I plan on passing this information on to the volunteers at the Computer History Museum! (I'm a member of the PDP-1 Restoration Team). I will, of course, give you credit for bringing it to my attention! --snip-- Cheers, Lyle Bickley Consulting West Inc. http://bickleywest.com "Black holes are where God is dividing by zero" From tivo.overo at gmail.com Sat Nov 17 15:14:24 2012 From: tivo.overo at gmail.com (Gary Lee Phillips) Date: Sat, 17 Nov 2012 14:14:24 -0600 Subject: [Simh] VAX XQ and wireless networking Message-ID: Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the VAX with OpenVMS 7.3. It works great, except for networking. The hardware is ASUS EeePC with Atheros AR5001 wireless network adapter and Atheros L2 Fast Ethernet adapter. I have libpcap version 1.1.1 (which is also callable as version 0.8) installed on Linux. The VAX module was compiled with networking enabled. The configuration file has an "attach xq0 wlan0" or "attach xq0 eth0" depending on which interface I'm trying to use. With the wired ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With the wireless (wlan0) neither one works. I've tried disabling the DECNET IV compatibility mode, which was recommended in one source for wireless compatibility, and also tried adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of the wireless adapter) to the configuration. The first did nothing I could discern, and the second generated an error message during startup of the SIMH module, saying that the MAC address was already in use and could not be assigned. Obviously the workaround is to use a cabled connection when the VAX emulator is needed, but this isn't very practical in most cases. Is there any other possibility I've overlooked? I notice that this is a frequently found question when using Google to look for answers, but I found no responses other than the two solutions mentioned above. I suppose a different network card emulation for VAX could make the wi-fi adapter look like a normal ethernet interface, but apparently that hasn't been done yet and it's probably beyond my very rudimentary C coding abilities. From dgahling at hotmail.com Sat Nov 17 15:24:02 2012 From: dgahling at hotmail.com (Dan Gahlinger) Date: Sat, 17 Nov 2012 15:24:02 -0500 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: Message-ID: It works...use the TUN interface and the bridging, works just fine. Dan. > Date: Sat, 17 Nov 2012 14:14:24 -0600 > From: tivo.overo at gmail.com > To: simh at trailing-edge.com > Subject: [Simh] VAX XQ and wireless networking > > Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the > VAX with OpenVMS 7.3. It works great, except for networking. The > hardware is ASUS EeePC with Atheros AR5001 wireless network adapter > and Atheros L2 Fast Ethernet adapter. > > I have libpcap version 1.1.1 (which is also callable as version 0.8) > installed on Linux. The VAX module was compiled with networking > enabled. > > The configuration file has an "attach xq0 wlan0" or "attach xq0 eth0" > depending on which interface I'm trying to use. With the wired > ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With the > wireless (wlan0) neither one works. > > I've tried disabling the DECNET IV compatibility mode, which was > recommended in one source for wireless compatibility, and also tried > adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of the > wireless adapter) to the configuration. The first did nothing I could > discern, and the second generated an error message during startup of > the SIMH module, saying that the MAC address was already in use and > could not be assigned. > > Obviously the workaround is to use a cabled connection when the VAX > emulator is needed, but this isn't very practical in most cases. Is > there any other possibility I've overlooked? > > I notice that this is a frequently found question when using Google to > look for answers, but I found no responses other than the two > solutions mentioned above. I suppose a different network card > emulation for VAX could make the wi-fi adapter look like a normal > ethernet interface, but apparently that hasn't been done yet and it's > probably beyond my very rudimentary C coding abilities. > _______________________________________________ > Simh mailing list > Simh at trailing-edge.com > http://mailman.trailing-edge.com/mailman/listinfo/simh -------------- next part -------------- An HTML attachment was scrubbed... URL: From bqt at softjar.se Sat Nov 17 15:57:46 2012 From: bqt at softjar.se (Johnny Billquist) Date: Sat, 17 Nov 2012 21:57:46 +0100 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: Message-ID: <50A7FA4A.9080607@softjar.se> On 2012-11-17 21:14, Gary Lee Phillips wrote: > Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the > VAX with OpenVMS 7.3. It works great, except for networking. The > hardware is ASUS EeePC with Atheros AR5001 wireless network adapter > and Atheros L2 Fast Ethernet adapter. > > I have libpcap version 1.1.1 (which is also callable as version 0.8) > installed on Linux. The VAX module was compiled with networking > enabled. > > The configuration file has an "attach xq0 wlan0" or "attach xq0 eth0" > depending on which interface I'm trying to use. With the wired > ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With the > wireless (wlan0) neither one works. > > I've tried disabling the DECNET IV compatibility mode, which was > recommended in one source for wireless compatibility, and also tried > adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of the > wireless adapter) to the configuration. The first did nothing I could > discern, and the second generated an error message during startup of > the SIMH module, saying that the MAC address was already in use and > could not be assigned. > > Obviously the workaround is to use a cabled connection when the VAX > emulator is needed, but this isn't very practical in most cases. Is > there any other possibility I've overlooked? > > I notice that this is a frequently found question when using Google to > look for answers, but I found no responses other than the two > solutions mentioned above. I suppose a different network card > emulation for VAX could make the wi-fi adapter look like a normal > ethernet interface, but apparently that hasn't been done yet and it's > probably beyond my very rudimentary C coding abilities. The "problem" is that a wlan interface is not really the same thing as an ethernet interface. As Dan suggested (I noticed), use a tun interface, and let than tunnel into the wlan instead, and things might work better. Johnny -- Johnny Billquist || "I'm on a bus || on a psychedelic trip email: bqt at softjar.se || Reading murder books pdp is alive! || tryin' to stay hip" - B. Idol From tivo.overo at gmail.com Sat Nov 17 15:58:13 2012 From: tivo.overo at gmail.com (Gary Lee Phillips) Date: Sat, 17 Nov 2012 14:58:13 -0600 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: Message-ID: I should have mentioned that I've tried setting up bridging and taptap without success when using the wireless interface. I use it on my desktop PCs (wired ethernet and Ubuntu) where it works OK though it's very cumbersome to start up and shut down. I thought the main advantage of a bridge was the fact that it lets the host and the emulated machine speak directly to one another via network packets, and that's a feature I don't really need in this case. Using the serial port emulation with a VT100 emulation over a tcp/ip socket is adequate and works well. On the desktop machines, the taptap module seems to have a tendency to "hang" on shutdown, even though the SIMH emulation has been closed down normally. The unreleased bridge port will keep the entire host from shutting down and requires a forced power off, which isn't very nice. On 11/17/12, Dan Gahlinger wrote: > > It works...use the TUN interface and the bridging, works just fine. > Dan. > >> Date: Sat, 17 Nov 2012 14:14:24 -0600 >> From: tivo.overo at gmail.com >> To: simh at trailing-edge.com >> Subject: [Simh] VAX XQ and wireless networking >> >> Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the >> VAX with OpenVMS 7.3. It works great, except for networking. The >> hardware is ASUS EeePC with Atheros AR5001 wireless network adapter >> and Atheros L2 Fast Ethernet adapter. >> >> I have libpcap version 1.1.1 (which is also callable as version 0.8) >> installed on Linux. The VAX module was compiled with networking >> enabled. >> >> The configuration file has an "attach xq0 wlan0" or "attach xq0 eth0" >> depending on which interface I'm trying to use. With the wired >> ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With the >> wireless (wlan0) neither one works. >> >> I've tried disabling the DECNET IV compatibility mode, which was >> recommended in one source for wireless compatibility, and also tried >> adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of the >> wireless adapter) to the configuration. The first did nothing I could >> discern, and the second generated an error message during startup of >> the SIMH module, saying that the MAC address was already in use and >> could not be assigned. >> >> Obviously the workaround is to use a cabled connection when the VAX >> emulator is needed, but this isn't very practical in most cases. Is >> there any other possibility I've overlooked? >> >> I notice that this is a frequently found question when using Google to >> look for answers, but I found no responses other than the two >> solutions mentioned above. I suppose a different network card >> emulation for VAX could make the wi-fi adapter look like a normal >> ethernet interface, but apparently that hasn't been done yet and it's >> probably beyond my very rudimentary C coding abilities. >> _______________________________________________ >> Simh mailing list >> Simh at trailing-edge.com >> http://mailman.trailing-edge.com/mailman/listinfo/simh > From dgahling at hotmail.com Sat Nov 17 16:03:56 2012 From: dgahling at hotmail.com (Dan Gahlinger) Date: Sat, 17 Nov 2012 16:03:56 -0500 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: , , Message-ID: actually that's exactly what you DO need to do. I do this myself with a wireless wlan interface and it works just fine. you need the bridge and attach to the TUN/TAP interface.and as the other post mentioned bridge the wlan interface the unreleased bridge port certainly will not keep the entire host from shutting down,I have no idea where you got such a crazy idea. it works just fine for the two systems I run this on wireless. I think this could even be made to work niceless on a raspberry pi rig. imagine that a vax-cluster of wireless embedded systems, very cool. > Date: Sat, 17 Nov 2012 14:58:13 -0600 > Subject: Re: [Simh] VAX XQ and wireless networking > From: tivo.overo at gmail.com > To: dgahling at hotmail.com > CC: simh at trailing-edge.com > > I should have mentioned that I've tried setting up bridging and taptap > without success when using the wireless interface. I use it on my > desktop PCs (wired ethernet and Ubuntu) where it works OK though it's > very cumbersome to start up and shut down. > > I thought the main advantage of a bridge was the fact that it lets the > host and the emulated machine speak directly to one another via > network packets, and that's a feature I don't really need in this > case. Using the serial port emulation with a VT100 emulation over a > tcp/ip socket is adequate and works well. > > On the desktop machines, the taptap module seems to have a tendency to > "hang" on shutdown, even though the SIMH emulation has been closed > down normally. The unreleased bridge port will keep the entire host > from shutting down and requires a forced power off, which isn't very > nice. > > On 11/17/12, Dan Gahlinger wrote: > > > > It works...use the TUN interface and the bridging, works just fine. > > Dan. > > > >> Date: Sat, 17 Nov 2012 14:14:24 -0600 > >> From: tivo.overo at gmail.com > >> To: simh at trailing-edge.com > >> Subject: [Simh] VAX XQ and wireless networking > >> > >> Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the > >> VAX with OpenVMS 7.3. It works great, except for networking. The > >> hardware is ASUS EeePC with Atheros AR5001 wireless network adapter > >> and Atheros L2 Fast Ethernet adapter. > >> > >> I have libpcap version 1.1.1 (which is also callable as version 0.8) > >> installed on Linux. The VAX module was compiled with networking > >> enabled. > >> > >> The configuration file has an "attach xq0 wlan0" or "attach xq0 eth0" > >> depending on which interface I'm trying to use. With the wired > >> ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With the > >> wireless (wlan0) neither one works. > >> > >> I've tried disabling the DECNET IV compatibility mode, which was > >> recommended in one source for wireless compatibility, and also tried > >> adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of the > >> wireless adapter) to the configuration. The first did nothing I could > >> discern, and the second generated an error message during startup of > >> the SIMH module, saying that the MAC address was already in use and > >> could not be assigned. > >> > >> Obviously the workaround is to use a cabled connection when the VAX > >> emulator is needed, but this isn't very practical in most cases. Is > >> there any other possibility I've overlooked? > >> > >> I notice that this is a frequently found question when using Google to > >> look for answers, but I found no responses other than the two > >> solutions mentioned above. I suppose a different network card > >> emulation for VAX could make the wi-fi adapter look like a normal > >> ethernet interface, but apparently that hasn't been done yet and it's > >> probably beyond my very rudimentary C coding abilities. > >> _______________________________________________ > >> Simh mailing list > >> Simh at trailing-edge.com > >> http://mailman.trailing-edge.com/mailman/listinfo/simh > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From oboguev at yahoo.com Sat Nov 17 16:04:42 2012 From: oboguev at yahoo.com (Sergey Oboguev) Date: Sat, 17 Nov 2012 13:04:42 -0800 (PST) Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: Message-ID: <1353186282.39615.YahooMailRC@web184301.mail.ne1.yahoo.com> I have not personally tried it, but here are few general remarks: SIMH DEQNA/DELQA virtual devices talk 802.3, whereas WiFi (802.11) is different from 802.3 in frame structure and everything else. WiFi may superficially look like Ethernet (i.e. have similarly looking MAC address), but it is actually not, so naive binding of XQ to host WiFi interface would not work for that reason. In addition, guest operating system software using DEQNA/DELQA will likewise be expecting 802.3 structured frames and protocol capabilities. Therefore, some form of translation between 802.3 and 802.11 or tunneling of 802.3 packets through 802.11 (or generic TCP/IP) would be required. If all you want is to connect SIMH instance to exterior TCP/IP network via IP connectivity only (i.e. do not need VAXcluster, LAT or DECnet), then regular host OS level bridging of tap interface (or Windows interface such as loopback or wired Ethernet interfaces) to WiFi interface would be one thing to try. Bridging is likely to suffice to translate IP packets from 802.3 MAC envelope to 802.11 and vice versa; that is it's purpose anyway. If, on the other hand, you want to interconnect multiple SIMH instances over WiFi utilizing not only IP, but also other Ethernet based protocols such as VAXcluster or LAT or DECnet, host OS bridging may fail to translate those protocols. In the latter case you'd need to find a way to tunnel 802.3 packets over 802.11 or over generic TCP/IP. General approach would be to attach SIMH XQ to Linux tap device on one side and hook tap/tun's other side into tunneling software. VDE (http://wiki.virtualsquare.org/wiki/index.php/VDE) does just that, but I personally do not have experience using VDE. Furthermore, there is some direct VDE support in the latest version of SIMH (assuming the simulator executable is built with VDE option enabled). I have not personally tried using it, but at a glance at the code, it looks like it allows to connect XQ to VDE_SWITCH directly (using vde:port syntax), not requiring tap device as an intermediary. -------------- next part -------------- An HTML attachment was scrubbed... URL: From tivo.overo at gmail.com Sat Nov 17 16:31:50 2012 From: tivo.overo at gmail.com (Gary Lee Phillips) Date: Sat, 17 Nov 2012 15:31:50 -0600 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: Message-ID: Might be some peculiarity of Ubuntu or my two Dell desktop PCs, but the system logs do in fact show that they are waiting for the tun0 device to close down and that it is "busy". I've waited as long as 30 minutes after the SIMH module was closed, and still no shutdown for Linux. The desktop machines have SIMH 3.8 rather than 3.9, which might also make some difference. I will try the bridge again, but previous attempts on the laptop failed to work for either the host or the guest system, using the same sequence of steps to set up the bridge that work with the desktop machines and wired ethernet. On 11/17/12, Dan Gahlinger wrote: > > actually that's exactly what you DO need to do. > I do this myself with a wireless wlan interface and it works just fine. > you need the bridge and attach to the TUN/TAP interface.and as the other > post mentioned bridge the wlan interface > the unreleased bridge port certainly will not keep the entire host from > shutting down,I have no idea where you got such a crazy idea. > it works just fine for the two systems I run this on wireless. > I think this could even be made to work niceless on a raspberry pi rig. > imagine that a vax-cluster of wireless embedded systems, very cool. From jg at jordi.guillaumes.name Sat Nov 17 16:46:40 2012 From: jg at jordi.guillaumes.name (Jordi Guillaumes i Pons) Date: Sat, 17 Nov 2012 22:46:40 +0100 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: <1353186282.39615.YahooMailRC@web184301.mail.ne1.yahoo.com> References: <1353186282.39615.YahooMailRC@web184301.mail.ne1.yahoo.com> Message-ID: <50A805C0.7000209@jordi.guillaumes.name> Al 17/11/12 22:04, En/na Sergey Oboguev ha escrit: > In the latter case you'd need to find a way to tunnel 802.3 packets > over 802.11 or over generic TCP/IP. General approach would be to > attach SIMH XQ to Linux tap device on one side and hook tap/tun's > other side into tunneling software. VDE > (http://wiki.virtualsquare.org/wiki/index.php/VDE) does just that, but > I personally do not have experience using VDE. Furthermore, there is > some direct VDE support in the latest version of SIMH (assuming the > simulator executable is built with VDE option enabled). I have not > personally tried using it, but at a glance at the code, it looks like > it allows to connect XQ to VDE_SWITCH directly (using vde:port > syntax), not requiring tap device as an intermediary. It does, and it works, but there is a caveat. I have not been successful bridging the tapx device which the vde switch exposes to the wlan physical interface. It is easily done with a real, wired ethernet, but not with a wlan. There is a solution. VDE comes with some support utilities which allow to "plug" into the virtual switch. Using vde_plug and dpipe (another VDE utility which provides two-way redirection between two unix processes) you can establish a ssh tunnel which acts as a transport pipe for the raw ethernet packages used by VDE. It works at level 2, so it should provide DECNET, LAT and LAVC connectivity (I have not tried the last one). As a bonus, it works with any IP link, including a tethered 3D link or even a POTS line. From simh at tj.merritts.org Sat Nov 17 18:07:06 2012 From: simh at tj.merritts.org (TJ Merritt) Date: Sat, 17 Nov 2012 15:07:06 -0800 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: Message-ID: <1353193626.2301.3.camel@vb-tjm> The TUN interface will get the packets injected into the system, but will not get them bridged across the wireless network which uses the MAC address for managing the wireless network. Since your MAC address for the virtual interface is not the same as the mac address for the host, reply packets over wireless will get dropped. The trick is to get the packets to go out with the host's MAC address and replies routed back to the virtual interface. VirtualBox does this by sending packets out with the original IP address but the MAC address of the host, they then record the virtual interface that the IP address came from, when the reply packet comes back, they use the IP address to look up the virtual interface associated with that IP address and route the packet to that interface. There are special cases of course for things like broadcast, multicast, and IP addresses that are shared between multiple VM's. VirtualBox does not currently support IPv6 addresses though. This would also not help for DECnet. This is effectively a layer 3 switching capability that works around the lack of layer 2 switching in the wireless network. Another approach would be to setup NAT so that the NAT layer can take care of the MAC address rewriting. It will also rewrite the IP address though. If this is satisfactory, you there should be plenty of NAT resources available to you on Mint such that no code writing is required. Adding virutal networking support for wireless networks would be a major undertaking for SimH, but if done, could eliminate most of the hiccups with virtual networking over wireless networks. Of course, not of these issues occur over wired netowrks. -- TJ On Sat, 2012-11-17 at 15:24 -0500, Dan Gahlinger wrote: > It works... > use the TUN interface and the bridging, works just fine. > > > Dan. > > > > Date: Sat, 17 Nov 2012 14:14:24 -0600 > > From: tivo.overo at gmail.com > > To: simh at trailing-edge.com > > Subject: [Simh] VAX XQ and wireless networking > > > > Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the > > VAX with OpenVMS 7.3. It works great, except for networking. The > > hardware is ASUS EeePC with Atheros AR5001 wireless network adapter > > and Atheros L2 Fast Ethernet adapter. > > > > I have libpcap version 1.1.1 (which is also callable as version 0.8) > > installed on Linux. The VAX module was compiled with networking > > enabled. > > > > The configuration file has an "attach xq0 wlan0" or "attach xq0 > eth0" > > depending on which interface I'm trying to use. With the wired > > ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With > the > > wireless (wlan0) neither one works. > > > > I've tried disabling the DECNET IV compatibility mode, which was > > recommended in one source for wireless compatibility, and also tried > > adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of > the > > wireless adapter) to the configuration. The first did nothing I > could > > discern, and the second generated an error message during startup of > > the SIMH module, saying that the MAC address was already in use and > > could not be assigned. > > > > Obviously the workaround is to use a cabled connection when the VAX > > emulator is needed, but this isn't very practical in most cases. Is > > there any other possibility I've overlooked? > > > > I notice that this is a frequently found question when using Google > to > > look for answers, but I found no responses other than the two > > solutions mentioned above. I suppose a different network card > > emulation for VAX could make the wi-fi adapter look like a normal > > ethernet interface, but apparently that hasn't been done yet and > it's > > probably beyond my very rudimentary C coding abilities. > > _______________________________________________ > > Simh mailing list > > Simh at trailing-edge.com > > http://mailman.trailing-edge.com/mailman/listinfo/simh > > _______________________________________________ > Simh mailing list > Simh at trailing-edge.com > http://mailman.trailing-edge.com/mailman/listinfo/simh From bqt at softjar.se Sat Nov 17 18:54:10 2012 From: bqt at softjar.se (Johnny Billquist) Date: Sun, 18 Nov 2012 00:54:10 +0100 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: <1353193626.2301.3.camel@vb-tjm> References: <1353193626.2301.3.camel@vb-tjm> Message-ID: <50A823A2.1050602@softjar.se> If that's how VirtualBox works, then it is rather broken. How about when you talk DECnet? There are no IP addresses involved at all. Johnny On 2012-11-18 00:07, TJ Merritt wrote: > The TUN interface will get the packets injected into the system, but > will not get them bridged across the wireless network which uses the MAC > address for managing the wireless network. Since your MAC address for > the virtual interface is not the same as the mac address for the host, > reply packets over wireless will get dropped. The trick is to get the > packets to go out with the host's MAC address and replies routed back to > the virtual interface. > > VirtualBox does this by sending packets out with the original IP address > but the MAC address of the host, they then record the virtual interface > that the IP address came from, when the reply packet comes back, they > use the IP address to look up the virtual interface associated with that > IP address and route the packet to that interface. There are special > cases of course for things like broadcast, multicast, and IP addresses > that are shared between multiple VM's. VirtualBox does not currently > support IPv6 addresses though. This would also not help for DECnet. > This is effectively a layer 3 switching capability that works around the > lack of layer 2 switching in the wireless network. > > Another approach would be to setup NAT so that the NAT layer can take > care of the MAC address rewriting. It will also rewrite the IP address > though. If this is satisfactory, you there should be plenty of NAT > resources available to you on Mint such that no code writing is > required. > > Adding virutal networking support for wireless networks would be a major > undertaking for SimH, but if done, could eliminate most of the hiccups > with virtual networking over wireless networks. > > Of course, not of these issues occur over wired netowrks. > > -- TJ > > On Sat, 2012-11-17 at 15:24 -0500, Dan Gahlinger wrote: >> It works... >> use the TUN interface and the bridging, works just fine. >> >> >> Dan. >> >> >>> Date: Sat, 17 Nov 2012 14:14:24 -0600 >>> From: tivo.overo at gmail.com >>> To: simh at trailing-edge.com >>> Subject: [Simh] VAX XQ and wireless networking >>> >>> Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the >>> VAX with OpenVMS 7.3. It works great, except for networking. The >>> hardware is ASUS EeePC with Atheros AR5001 wireless network adapter >>> and Atheros L2 Fast Ethernet adapter. >>> >>> I have libpcap version 1.1.1 (which is also callable as version 0.8) >>> installed on Linux. The VAX module was compiled with networking >>> enabled. >>> >>> The configuration file has an "attach xq0 wlan0" or "attach xq0 >> eth0" >>> depending on which interface I'm trying to use. With the wired >>> ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With >> the >>> wireless (wlan0) neither one works. >>> >>> I've tried disabling the DECNET IV compatibility mode, which was >>> recommended in one source for wireless compatibility, and also tried >>> adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of >> the >>> wireless adapter) to the configuration. The first did nothing I >> could >>> discern, and the second generated an error message during startup of >>> the SIMH module, saying that the MAC address was already in use and >>> could not be assigned. >>> >>> Obviously the workaround is to use a cabled connection when the VAX >>> emulator is needed, but this isn't very practical in most cases. Is >>> there any other possibility I've overlooked? >>> >>> I notice that this is a frequently found question when using Google >> to >>> look for answers, but I found no responses other than the two >>> solutions mentioned above. I suppose a different network card >>> emulation for VAX could make the wi-fi adapter look like a normal >>> ethernet interface, but apparently that hasn't been done yet and >> it's >>> probably beyond my very rudimentary C coding abilities. >>> _______________________________________________ >>> Simh mailing list >>> Simh at trailing-edge.com >>> http://mailman.trailing-edge.com/mailman/listinfo/simh >> >> _______________________________________________ >> Simh mailing list >> Simh at trailing-edge.com >> http://mailman.trailing-edge.com/mailman/listinfo/simh > > > _______________________________________________ > Simh mailing list > Simh at trailing-edge.com > http://mailman.trailing-edge.com/mailman/listinfo/simh > -- Johnny Billquist || "I'm on a bus || on a psychedelic trip email: bqt at softjar.se || Reading murder books pdp is alive! || tryin' to stay hip" - B. Idol From tivo.overo at gmail.com Sat Nov 17 19:02:48 2012 From: tivo.overo at gmail.com (Gary Lee Phillips) Date: Sat, 17 Nov 2012 18:02:48 -0600 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: <1353193626.2301.3.camel@vb-tjm> References: <1353193626.2301.3.camel@vb-tjm> Message-ID: Well, just to confirm what I thought I remembered: I tried again to set up bridging, and it is not possible using the method that works for wired ethernet. That is, I define a bridge br0 using the brctl command (as root of course) and then add interfaces to the bridge. I can add eth0 without difficulty, just as on other machines. But if I try to add the wireless interface wlan0, the command is rejected: can't add wlan0 to bridge br0: Operation not supported I have used iptables (and earlier versions of the command, ipchains or ifwadm) to set up software firewalls before, and even to NAT ip addresses, but I've never seen a way to make it change a MAC address. Even if I had, I have no idea how to apply that to the output of the SIMH XQ device. Are you saying I should create a tun device or something similar to which XQ0 can be attached, and then re-route packets from that device to wlan0 by using iptables? I believe if IP were working, then DECNET can be routed or encapsulated over IP. But in this particular case, DECNET is inessential if I have working IP. The working IP tools and clients are sufficient. And I thought DECNET Plus didn't need to mess with MAC addresses anyway. Isn't it just DECNET IV that does that? On 11/17/12, TJ Merritt wrote: > The TUN interface will get the packets injected into the system, but > will not get them bridged across the wireless network which uses the MAC > address for managing the wireless network. Since your MAC address for > the virtual interface is not the same as the mac address for the host, > reply packets over wireless will get dropped. The trick is to get the > packets to go out with the host's MAC address and replies routed back to > the virtual interface. > > VirtualBox does this by sending packets out with the original IP address > but the MAC address of the host, they then record the virtual interface > that the IP address came from, when the reply packet comes back, they > use the IP address to look up the virtual interface associated with that > IP address and route the packet to that interface. There are special > cases of course for things like broadcast, multicast, and IP addresses > that are shared between multiple VM's. VirtualBox does not currently > support IPv6 addresses though. This would also not help for DECnet. > This is effectively a layer 3 switching capability that works around the > lack of layer 2 switching in the wireless network. > > Another approach would be to setup NAT so that the NAT layer can take > care of the MAC address rewriting. It will also rewrite the IP address > though. If this is satisfactory, you there should be plenty of NAT > resources available to you on Mint such that no code writing is > required. > > Adding virutal networking support for wireless networks would be a major > undertaking for SimH, but if done, could eliminate most of the hiccups > with virtual networking over wireless networks. > > Of course, not of these issues occur over wired netowrks. > > -- TJ > > On Sat, 2012-11-17 at 15:24 -0500, Dan Gahlinger wrote: >> It works... >> use the TUN interface and the bridging, works just fine. >> >> >> Dan. >> >> >> > Date: Sat, 17 Nov 2012 14:14:24 -0600 >> > From: tivo.overo at gmail.com >> > To: simh at trailing-edge.com >> > Subject: [Simh] VAX XQ and wireless networking >> > >> > Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the >> > VAX with OpenVMS 7.3. It works great, except for networking. The >> > hardware is ASUS EeePC with Atheros AR5001 wireless network adapter >> > and Atheros L2 Fast Ethernet adapter. >> > >> > I have libpcap version 1.1.1 (which is also callable as version 0.8) >> > installed on Linux. The VAX module was compiled with networking >> > enabled. >> > >> > The configuration file has an "attach xq0 wlan0" or "attach xq0 >> eth0" >> > depending on which interface I'm trying to use. With the wired >> > ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With >> the >> > wireless (wlan0) neither one works. >> > >> > I've tried disabling the DECNET IV compatibility mode, which was >> > recommended in one source for wireless compatibility, and also tried >> > adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of >> the >> > wireless adapter) to the configuration. The first did nothing I >> could >> > discern, and the second generated an error message during startup of >> > the SIMH module, saying that the MAC address was already in use and >> > could not be assigned. >> > >> > Obviously the workaround is to use a cabled connection when the VAX >> > emulator is needed, but this isn't very practical in most cases. Is >> > there any other possibility I've overlooked? >> > >> > I notice that this is a frequently found question when using Google >> to >> > look for answers, but I found no responses other than the two >> > solutions mentioned above. I suppose a different network card >> > emulation for VAX could make the wi-fi adapter look like a normal >> > ethernet interface, but apparently that hasn't been done yet and >> it's >> > probably beyond my very rudimentary C coding abilities. >> > _______________________________________________ >> > Simh mailing list >> > Simh at trailing-edge.com >> > http://mailman.trailing-edge.com/mailman/listinfo/simh >> >> _______________________________________________ >> Simh mailing list >> Simh at trailing-edge.com >> http://mailman.trailing-edge.com/mailman/listinfo/simh > > > From simh at tj.merritts.org Sun Nov 18 10:10:21 2012 From: simh at tj.merritts.org (TJ Merritt) Date: Sun, 18 Nov 2012 07:10:21 -0800 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: <50A823A2.1050602@softjar.se> References: <1353193626.2301.3.camel@vb-tjm> <50A823A2.1050602@softjar.se> Message-ID: <1353251421.2301.6.camel@vb-tjm> It's not so much that VirtualBox is broken, but rather it is working around the brokenness in 802.11. VirtualBox has hooks for supporting L3 bridging of any L3 protocol, but there is only an implementation for IPv4. I'm aware of this issue only because of the annoyance of IPv6 bridging failing on wireless on VirtualBox VM's where IPv4 works. Overall the networking support in Virtual Box is quite superior. -- TJ On Sun, 2012-11-18 at 00:54 +0100, Johnny Billquist wrote: > If that's how VirtualBox works, then it is rather broken. > How about when you talk DECnet? There are no IP addresses involved at all. > > Johnny > From jg at jordi.guillaumes.name Mon Nov 19 12:16:17 2012 From: jg at jordi.guillaumes.name (Jordi Guillaumes i Pons) Date: Mon, 19 Nov 2012 18:16:17 +0100 Subject: [Simh] Possible problem in Extra-VAXen branch: Message-ID: <50AA6961.6090404@jordi.guillaumes.name> vax780 seems to be broken since a few days ago: jguillaumes at pi01 ~/emul/vax780 $ ../simh/BIN/vax780 vax780.ini VAX780 simulator V4.0-0 libpcap version 1.3.0 Eth: opened OS device vde:/tmp/vde.ctl Listening on port 32123 (socket 8) Modem control activated Auto disconnect activated sim> b rp Loading boot code from vmb.exe Saving boot code to vmb.exe Loading boot code from vmb.exe Address space exceeded, PC: 00000000 (HALT) Backing to the Oct 31 commit (69666f1480d494c225a1e82b8b38568792948382): jguillaumes at pi01 ~/emul/vax780 $ ../simh/BIN/vax780 vax780.ini VAX780 simulator V4.0-0 libpcap version 1.3.0 Eth: opened OS device vde:/tmp/vde.ctl Listening on port 32123 (socket 8) Modem control activated Auto disconnect activated sim> b rp Loading boot code from vmb.exe VAX/VMS Version V4.7 28-Oct-1987 13:00 -------------- next part -------------- An HTML attachment was scrubbed... URL: From petermallan at gmail.com Wed Nov 21 15:27:27 2012 From: petermallan at gmail.com (Peter Allan) Date: Wed, 21 Nov 2012 20:27:27 +0000 Subject: [Simh] Adding a third massbus to the VAX780 emulator Message-ID: Many, many years ago, I used to be the system manager of a VAX-11/780 system that was part of the Starlink astronomical data processing network in the UK. One of my pet projects is to have an emulation of this as close to the original as I can get it. The system had two RP06 disk drives and two System Industries drives that emulated RM05 drives. Since the System Industries drives interfaced to the 780 through their own massbus adapter, the RP06 drives appeared as devices DBA0 and DBA1, and the emulated RM05 drives appeared as DRA0 and DRA1. I can create a similar system using simh, but with all of the drives on a single massbus, the drives appear as devices DBA0, DBA1, DRA2 and DRA3. This apparently trivial difference annoys me so I am looking for a way of having both the RP06 and RM05 drives appear as units 0 and 1. I presume that this will require modifying simh to add a third massbus adapter. (Please do tell me if there is an easier way.) Having looked at the simh code, I am reasonably confident of being able to add a third massbus. However, how can I attach RP drives to one massbus and RM drives to another? Both drives are handled by the same function, i.e. pdp11_rp.c. My thoughts at the moment are that I should copy pdp11_rp.c to a new file pdp11_rm.c and remove the code that handles the RM disk from pdp11_rp.c and vice versa for the RP code in pdp11_rm.c I can see that there is more work to do in terms of splitting the handing of interrupts and the like, but before diving in, does my basic strategy seem correct? It would be more like real life (I think) to be able to add RM and RP drives to both massbuses that handle disks, but I don't see how to do that without a lot more work. And before anyone says it, I know I am mad to be trying to do this. The effort is clearly disproportionate to the benefit, but it is a hobby after all. All comments gratefully received. Peter Allan, petermallan at gmail.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From bqt at softjar.se Wed Nov 21 16:03:17 2012 From: bqt at softjar.se (Johnny Billquist) Date: Wed, 21 Nov 2012 22:03:17 +0100 Subject: [Simh] Adding a third massbus to the VAX780 emulator In-Reply-To: References: Message-ID: <50AD4195.5000401@softjar.se> I think this is totally the wrong strategy. If anything, you should try to model this as close to how the actual hardware worked, and not do some special thing that happens to just match your very specific configuration. What you need is to split the massbus handling, as such, from the specific disks. You should then be able to create massbusses for you machine, and then you attach disks to the massbus. It is a hierarchy, and the config should reflect that. I think simh's current approach of presenting things in a very flat way is a bit too simplistic. I got annoyed the other day when I wanted more than one DZ11 on a configuration, and realized I could not configure the controllers as such at all, but instead configured the total number of serial lines, with an implicit overflow from one controller to the next. In real life, you might have a machine with four serial lines, but all attached to separate DZ11. Just specifying the endpoints without the ability to specify the configuration in the middle layers will cause issues like the ones you have, because simh then makes assumptions for you. (Such as having just one massbus, and all disks attached to that one.) In fact, I wonder if this will actually totally break something like RSX-11M, which do not allow mixed device massbuses. Ie. in 11M you cannot have both RM and RP disks on the same massbus. M+ allows this, but then, M+ is rather more advanced... :-) Johnny On 2012-11-21 21:27, Peter Allan wrote: > Many, many years ago, I used to be the system manager of a VAX-11/780 > system that was part of the Starlink astronomical data processing > network in the UK. One of my pet projects is to have an emulation of > this as close to the original as I can get it. The system had two RP06 > disk drives and two System Industries drives that emulated RM05 drives. > Since the System Industries drives interfaced to the 780 through their > own massbus adapter, the RP06 drives appeared as devices DBA0 and DBA1, > and the emulated RM05 drives appeared as DRA0 and DRA1. > I can create a similar system using simh, but with all of the drives on > a single massbus, the drives appear as devices DBA0, DBA1, DRA2 and > DRA3. This apparently trivial difference annoys me so I am looking for a > way of having both the RP06 and RM05 drives appear as units 0 and 1. > I presume that this will require modifying simh to add a third massbus > adapter. (Please do tell me if there is an easier way.) > Having looked at the simh code, I am reasonably confident of being able > to add a third massbus. However, how can I attach RP drives to one > massbus and RM drives to another? Both drives are handled by the same > function, i.e. pdp11_rp.c. My thoughts at the moment are that I should > copy pdp11_rp.c to a new file pdp11_rm.c and remove the code that > handles the RM disk from pdp11_rp.c and vice versa for the RP code in > pdp11_rm.c > I can see that there is more work to do in terms of splitting the > handing of interrupts and the like, but before diving in, does my basic > strategy seem correct? It would be more like real life (I think) to be > able to add RM and RP drives to both massbuses that handle disks, but I > don't see how to do that without a lot more work. > And before anyone says it, I know I am mad to be trying to do this. The > effort is clearly disproportionate to the benefit, but it is a hobby > after all. > All comments gratefully received. > Peter Allan, petermallan at gmail.com > > > _______________________________________________ > Simh mailing list > Simh at trailing-edge.com > http://mailman.trailing-edge.com/mailman/listinfo/simh > From petermallan at gmail.com Thu Nov 22 04:36:18 2012 From: petermallan at gmail.com (Peter Allan) Date: Thu, 22 Nov 2012 09:36:18 +0000 Subject: [Simh] Adding a third massbus to the VAX780 emulator In-Reply-To: <0CC6789C1C831B4C8CCFF49D45D7010F840266F4B1@REDROOF2.alohasunset.com> References: <0CC6789C1C831B4C8CCFF49D45D7010F840266F4B1@REDROOF2.alohasunset.com> Message-ID: Thanks to all for your comments. Kevin, yes I could define logical names that would let me refer to the drives using the names DRA0 and DRA1, but they would still show up as DRA2 and DRA3 on the output of "sho dev d" for example. I would like to do better if I can. Johnny, I agree in principle, but I really don't want to get into rewriting too much of SIMH. Mark, that sounds like a very useful suggestion and I will look into how the RQ controllers are implemented. Hopefully it will indeed let me add any drive to any massbus, which is as it should be. If I succeed, it should be generally useful. My original suggestion was clearly a "dirty hack" and even if it worked, would not be suitable for airing in public. Cheers Peter Allan On 21 November 2012 22:09, Mark Pizzolato - Info Comm wrote: > Hi Peter,**** > > ** ** > > I have no specific knowledge of how multiple mass busses may be ?wired? > into simh. However, once you figure that detail out. I suggest that you > consider the idea of extending the existing pdp11_rp.c to support multiple > ?device? instances where each device is attached to a desired massbus and > has its own set of units. You may want to look at how pdp11_rq.c supports > multiple RQ instances as an example. Meanwhile, if you?d ever like to have > your changes make it back into the general simh code base, then please > start from the latest pdp11_rp.c which can be found at: > https://github.com/simh/simh/blob/Extra-VAXen/PDP11/pdp11_rp.c and a zip > of the current Extra-VAXen branch at: > https://github.com/simh/simh/archive/Extra-VAXen.zip**** > > I am in the middle of tracking down a bug in this code right now ( > https://github.com/simh/simh/issues/14), so feel free to start looking at > things now, but be prepared to take an update to pdp11_rp.c when I resolve > the issue.**** > > ** ** > > Good Luck,**** > > ** ** > > **- **Mark Pizzolato**** > > ** ** > > *From:* simh-bounces at trailing-edge.com [mailto: > simh-bounces at trailing-edge.com] *On Behalf Of *Peter Allan > *Sent:* Wednesday, November 21, 2012 12:27 PM > *To:* simh at trailing-edge.com > *Subject:* [Simh] Adding a third massbus to the VAX780 emulator**** > > ** ** > > Many, many years ago, I used to be the system manager of a VAX-11/780 > system that was part of the Starlink astronomical data processing network > in the UK. One of my pet projects is to have an emet defulation of this > as close to the original as I can get it. The system had two RP06 disk > drives and two System Industries drives that emulated RM05 drives. Since > the System Industries drives interfaced to the 780 through their own > massbus adapter, the RP06 drives appeared as devices DBA0 and DBA1, and the > emulated RM05 drives appeared as DRA0 and DRA1.**** > > **** > > I can create a similar system using simh, but with all of the drives on a > single massbus, the drives appear as devices DBA0, DBA1, DRA2 and DRA3. > This apparently trivial difference annoys me so I am looking for a way of > having both the RP06 and RM05 drives appear as units 0 and 1.**** > > **** > > I presume that this will require modifying simh to add a third massbus > adapter. (Please do tell me if there is an easier way.)**** > > **** > > Having looked at the simh code, I am reasonably confident of being able to > add a third massbus. However, how can I attach RP drives to one massbus and > RM drives to another? Both drives are handled by the same function, i.e. > pdp11_rp.c. My thoughts at the moment are that I should copy pdp11_rp.c to > a new file pdp11_rm.c and remove the code that handles the RM disk from > pdp11_rp.c and vice versa for the RP code in pdp11_rm.c**** > > **** > > I can see that there is more work to do in terms of splitting the handing > of interrupts and the like, but before diving in, does my basic strategy > seem correct? It would be more like real life (I think) to be able to add > RM and RP drives to both massbuses that handle disks, but I don't see how > to do that without a lot more work.**** > > **** > > And before anyone says it, I know I am mad to be trying to do this. The > effort is clearly disproportionate to the benefit, but it is a hobby after > all.**** > > **** > > All comments gratefully received.**** > > **** > > Peter Allan, petermallan at gmail.com**** > -------------- next part -------------- An HTML attachment was scrubbed... URL: From ddyer-simh at real-me.net Mon Nov 26 13:50:51 2012 From: ddyer-simh at real-me.net (Dave Dyer) Date: Mon, 26 Nov 2012 10:50:51 -0800 Subject: [Simh] even older PDP-10 macro Message-ID: <20121126185059.4DA491E0100@mini-me.trailing-edge.com> For my particular PDP-10 archeology, I need an even older version of Macro-10 than I can find in the various archive tapes. My recollection is that I need version 47, which would have been current about 1975. Either a binary or source would probably do. I hope someone has an image of a system tape from that era that was thought to be not important because there were newer tapes available). (back story; Either macro 50+ had new bugs or Dec deliberately changed the semantics of macros. In either case, a compile that had complicated macros no longer worked on macro version 50+, and even in the day I had to use "r old:macro" to compile.) From tshoppa at wmata.com Mon Nov 26 14:53:44 2012 From: tshoppa at wmata.com (Shoppa, Tim) Date: Mon, 26 Nov 2012 19:53:44 +0000 Subject: [Simh] even older PDP-10 macro In-Reply-To: <20121126185059.4DA491E0100@mini-me.trailing-edge.com> References: <20121126185059.4DA491E0100@mini-me.trailing-edge.com> Message-ID: <303A17BD5F8FA34DA45EEC245271AC0B2524B9C3@JGEX2K10MBX2.wmata.local> One big syntax change that happened between Macro V52 and Macro V53, was that support for "line starting with comma is a comment" was removed. The change that removed this support, seems to be reversible, i.e. a later Macro version could have this support added back in. I don't know when the change for semantics of complex macros was changed. Most of my efforts with early Macro sources, are related to compiling the PDP-6 monitor sources that are available, and it's been a long time since I've messed with that. Can't help with V47. But my notes on Macro versions from 50A onwards, including where they are available in the archives: 50A(441) 1-Jul-1976 As a Non-DEC update on DECUS 10-301 to add a user pushdown stack Patch between DEC and Non-DEC seems reversible. 53(1020) 27-Feb-1978 Has edit history from 50A through 53. TOPS-20 release 3 53A(1152) 19-Jul-1979 TOPS-20 release 3A 53B(1242) 3-Aug-1983 TOPS-20 release 6.1 53B(1244) 28-Jan-1986 March 86 TOPS-10 CUSP tapes 53B(1244) 13-Nov-85 TOPS-20 KS UPD 53B(1247) 29-Aug-86 Sept 88 TOPS-10 CUSP tapes 53B(1252) 24-Jun-88 TOPS-10 7.04 TSU01 -----Original Message----- From: simh-bounces at trailing-edge.com [mailto:simh-bounces at trailing-edge.com] On Behalf Of Dave Dyer Sent: Monday, November 26, 2012 1:51 PM To: Simh at trailing-edge.com Subject: [Simh] even older PDP-10 macro For my particular PDP-10 archeology, I need an even older version of Macro-10 than I can find in the various archive tapes. My recollection is that I need version 47, which would have been current about 1975. Either a binary or source would probably do. I hope someone has an image of a system tape from that era that was thought to be not important because there were newer tapes available). (back story; Either macro 50+ had new bugs or Dec deliberately changed the semantics of macros. In either case, a compile that had complicated macros no longer worked on macro version 50+, and even in the day I had to use "r old:macro" to compile.) _______________________________________________ Simh mailing list Simh at trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh From ddyer-simh at real-me.net Mon Nov 26 15:22:24 2012 From: ddyer-simh at real-me.net (Dave Dyer) Date: Mon, 26 Nov 2012 12:22:24 -0800 Subject: [Simh] even older PDP-10 macro Message-ID: <20121126202231.2309E1E0126@mini-me.trailing-edge.com> I suspect the change that broke later versions of macro was related to self redefining macros. The code in question used a complicated scheme to build a hashed symbol table. From robert.jarratt at ntlworld.com Sat Nov 3 18:15:51 2012 From: robert.jarratt at ntlworld.com (Rob Jarratt) Date: Sat, 3 Nov 2012 22:15:51 -0000 Subject: [Simh] Slightly OT: Headless SIMH on Raspberry Pi Message-ID: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> I want to run SIMH on my headless Raspberry Pi, to do so I use putty to connect to the pi and log in. From the shell I can run SIMH just fine. What I would like though is for my session not to disappear if I close putty and to be able to reconnect to the session at some later time, with SIMH still running. I am not really a Unix expert, is there a way to do this? Regards Rob -------------- next part -------------- An HTML attachment was scrubbed... URL: From abs at absd.org Sat Nov 3 18:41:30 2012 From: abs at absd.org (David Brownlee) Date: Sat, 3 Nov 2012 22:41:30 +0000 Subject: [Simh] Slightly OT: Headless SIMH on Raspberry Pi In-Reply-To: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> References: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> Message-ID: On 3 November 2012 22:15, Rob Jarratt wrote: > I want to run SIMH on my headless Raspberry Pi, to do so I use putty to > connect to the pi and log in. From the shell I can run SIMH just fine. What > I would like though is for my session not to disappear if I close putty and > to be able to reconnect to the session at some later time, with SIMH still > running. I am not really a Unix expert, is there a way to do this? "screen" http://www.gnu.org/software/screen/ or "tmux" http://tmux.sourceforge.net/ are two very useful tools for just this :) From md.benson at gmail.com Sat Nov 3 19:18:28 2012 From: md.benson at gmail.com (Mark Benson) Date: Sat, 3 Nov 2012 23:18:28 +0000 Subject: [Simh] Slightly OT: Headless SIMH on Raspberry Pi In-Reply-To: References: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> Message-ID: On 3 Nov 2012, at 22:41, David Brownlee wrote: > On 3 November 2012 22:15, Rob Jarratt wrote: >> I want to run SIMH on my headless Raspberry Pi, to do so I use putty to >> connect to the pi and log in. From the shell I can run SIMH just fine. What >> I would like though is for my session not to disappear if I close putty and >> to be able to reconnect to the session at some later time, with SIMH still >> running. I am not really a Unix expert, is there a way to do this? > > "screen" http://www.gnu.org/software/screen/ or "tmux" > http://tmux.sourceforge.net/ are two very useful tools for just this > :) I use gnu 'screen' extensively on my Rasbperry Pi VAX setups to do just that. Works a charm, and you can detach the session at any time and it will keep buffering. Right now I recommend this as the best way to run SimH headless and without a need for a perpetual terminal. I'm hoping something better might come along that doesn't require additional software, though. -- Mark Benson http://DECtec.info Twitter: @DECtecInfo HECnet: STAR69::MARK Online Resource & Mailing List for DEC Enthusiasts. From robert.jarratt at ntlworld.com Sun Nov 4 02:27:40 2012 From: robert.jarratt at ntlworld.com (Rob Jarratt) Date: Sun, 4 Nov 2012 07:27:40 -0000 Subject: [Simh] Slightly OT: Headless SIMH on Raspberry Pi In-Reply-To: References: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> Message-ID: <028101cdba5d$e1b7b3d0$a5271b70$@ntlworld.com> > -----Original Message----- > From: simh-bounces at trailing-edge.com [mailto:simh-bounces at trailing- > edge.com] On Behalf Of Mark Benson > Sent: 03 November 2012 23:18 > To: simh at trailing-edge.com > Subject: Re: [Simh] Slightly OT: Headless SIMH on Raspberry Pi > > > On 3 Nov 2012, at 22:41, David Brownlee wrote: > > > On 3 November 2012 22:15, Rob Jarratt > wrote: > >> I want to run SIMH on my headless Raspberry Pi, to do so I use putty > >> to connect to the pi and log in. From the shell I can run SIMH just > >> fine. What I would like though is for my session not to disappear if > >> I close putty and to be able to reconnect to the session at some > >> later time, with SIMH still running. I am not really a Unix expert, is there > a way to do this? > > > > "screen" http://www.gnu.org/software/screen/ or "tmux" > > http://tmux.sourceforge.net/ are two very useful tools for just this > > :) > > I use gnu 'screen' extensively on my Rasbperry Pi VAX setups to do just that. > Works a charm, and you can detach the session at any time and it will keep > buffering. > > Right now I recommend this as the best way to run SimH headless and > without a need for a perpetual terminal. I'm hoping something better might > come along that doesn't require additional software, though. > > -- > > Mark Benson > > http://DECtec.info > Twitter: @DECtecInfo > HECnet: STAR69::MARK > > Online Resource & Mailing List for DEC Enthusiasts. > > _______________________________________________ > Simh mailing list > Simh at trailing-edge.com > http://mailman.trailing-edge.com/mailman/listinfo/simh Thanks to everyone for the replies. I am trying out screen now and it seems to be working well. Regards Rob From paulhardy2 at btinternet.com Sun Nov 4 13:50:28 2012 From: paulhardy2 at btinternet.com (Paul Hardy 2) Date: Sun, 4 Nov 2012 18:50:28 -0000 Subject: [Simh] Slightly OT: Headless SIMH on Raspberry Pi In-Reply-To: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> References: <024d01cdba10$ca34e680$5e9eb380$@ntlworld.com> Message-ID: <000301cdbabd$422aaa10$c67ffe30$@btinternet.com> I use VNC to connect to my headless Raspberry Pi to run a SimH VAX- see (http://gettingstartedwithraspberrypi.tumblr.com/post/24142374137/setting-up -a-vnc-server). This survives disconnection OK. Paul. From: simh-bounces at trailing-edge.com [mailto:simh-bounces at trailing-edge.com] On Behalf Of Rob Jarratt Sent: 03 November 2012 22:16 To: simh at trailing-edge.com Subject: [Simh] Slightly OT: Headless SIMH on Raspberry Pi I want to run SIMH on my headless Raspberry Pi, to do so I use putty to connect to the pi and log in. From the shell I can run SIMH just fine. What I would like though is for my session not to disappear if I close putty and to be able to reconnect to the session at some later time, with SIMH still running. I am not really a Unix expert, is there a way to do this? Regards Rob -------------- next part -------------- An HTML attachment was scrubbed... URL: From oboguev at yahoo.com Sat Nov 10 15:38:55 2012 From: oboguev at yahoo.com (Sergey Oboguev) Date: Sat, 10 Nov 2012 12:38:55 -0800 (PST) Subject: [Simh] Multiprocessor VAX simulator In-Reply-To: <4CD39E08-92F1-44F5-A56C-21C92E2808A3@gmail.com> References: <4CD39E08-92F1-44F5-A56C-21C92E2808A3@gmail.com> Message-ID: <1352579935.10136.YahooMailRC@web184302.mail.ne1.yahoo.com> VAX MP is a derivative variant of SIMH VAX simulator capable to execute OpenVMS (VAX/VMS) in true SMP mode on multiprocessor host machines, including modern commodity multicore/hyperthreaded PCs and Macs. Initial release of VAX MP supports Intel x86 and x64 processors based machines as host hardware platform with Windows, Linux and OS X as host operating systems. The number of VAX MP virtual processors is limited to the number of host machine's simultaneous execution units (logical CPUs). For example, PC with quad-core i7 processor has hyperthreaded cores and would allow to run VAX MP instance with up to 8 virtual processors. VAX MP had been tested to OpenVMS VAX theoretical limit of 32 processors. VAX MP is an architectural simulator. The focus is on OpenVMS SMP capability, rather than simulation of specific historical hardware. Accordingly VAX MP does not attempt to simulate any historically existing VAX multiprocessor model, although it may be used as a stepping stone towards further development of such a simulator. Instead VAX MP simulates a multiprocessor variant of MicroVAX 3900, a machine that never historically existed, but complies generic VAX architecture and runs OpenVMS in SMP mode. VAX MP is currently downloadable from my pesonal web page that hosts source code, documentation, prebuilt binaries and screenshots. http://www.oboguev.net/vax_mp Those nostalgic for "real iron" may also view photographs of a session using VT320 connected via DECserver 200 to a VAXcluster of two VAX MP nodes, each with 8 processors, under heavy cluster IO traffic. Important legal notice: Although on a purely technical level VAX MP does not require VMS SMP extension license to run and OpenVMS runs fully functional in SMP mode on VAX MP without prompting for SMP extension license, however from the legal standpoint running VMS in SMP mode may require a license for VMS SMP extension. Whether such license is needed for hobbyist use is a gray area. Please be sure to consult "Legal information notice B" in "VAX MP OpenVMS User Manual" for more information regarding licensing situation. I personally do have SMP extension license legally sufficient to cover VAX MP development and testing "in the white", but your mileage may vary. I am unable to advise whether hobbyist use of OpenVMS VAX in multiprocessor mode would require SMP extension license. From oboguev at yahoo.com Sun Nov 11 16:09:17 2012 From: oboguev at yahoo.com (Sergey Oboguev) Date: Sun, 11 Nov 2012 13:09:17 -0800 (PST) Subject: [Simh] Virtual VAXen In-Reply-To: <1341840220.61205.YahooMailRC@web181119.mail.ne1.yahoo.com> References: <4FF9F720.2010207@supnik.org> <1341840220.61205.YahooMailRC@web181119.mail.ne1.yahoo.com> Message-ID: <1352668157.43838.YahooMailRC@web184305.mail.ne1.yahoo.com> I came across another article regarding the history of VAX virtualization project at DEC. Surprisingly, the projects dates back to 1981 and ran till 1990. The goal was to achieve A1 security level certification. Steve Lipner et al., "Lessons from VAX/SVS for High Assurance VM Systems" http://dx.doi.org/10.1109/MSP.2012.87 http://www.cse.psu.edu/~tjaeger/cse543-f12/docs/SP_SPSI-2012-03-0042 R1_Zurko.pdf ________________________________ From: Sergey Oboguev To: simh at trailing-edge.com Sent: Mon, July 9, 2012 6:23:41 AM Subject: Re: [Simh] Virtual VAXen > Ah yes, the VVAX option... another attempted graft on the VAX architecture >tree. Unlike vectors, it never got as far as a real implementation. According to VVAX / VAX Security Kernel developers, while it was never released externally as a product, there was internal proof-of-concept implementation on 11/730 and production quality implementation on 8800, the latter field-tested in 1989 at a number of external customers' sites. http://dx.doi.org/10.1109/32.106971 http://www.cse.psu.edu/~tjaeger/cse543-f06/papers/vax_vmm.pdf http://dx.doi.org/10.1109/RISP.1990.63834 http://www.scs.stanford.edu/nyu/02sp/sched/vmm.pdf And here is somewhat more detailed technical description: http://dx.doi.org/10.1109/ISCA.1991.1021630 http://mprc.pku.cn/mentors/training/ISCAreading/1991/p380-hall/p380-hall.pdf -------------- next part -------------- An HTML attachment was scrubbed... URL: From beebe at math.utah.edu Sat Nov 17 11:15:47 2012 From: beebe at math.utah.edu (Nelson H. F. Beebe) Date: Sat, 17 Nov 2012 09:15:47 -0700 (MST) Subject: [Simh] new papers on von Neumann's programs and machine Message-ID: [This posting is slightly off-target, because it is about virtual machines, and not Simh specifically. Nevertheless, the poster feels that it may well be of interest to list readers.] The latest issue of the IEEE journal Computer arrived in my mailbox a couple of days ago. It contains an interesting article that describes an emulator for John von Neumann's IAS (Institute for Advanced Study) computer, designed in the late 1940s, but not fully operational until 1952 (see, e.g., http://en.wikipedia.org/wiki/IAS_machine ). The article also describes the testing of several early programs written by von Neumann and his close collaborator, Herman Goldstine. The article authors report finding several bugs in the programs (all of which were published about five years before actual hardware became operational). Some of the errors arise from transcription from technical reports into the volumes of the Collected Works of John von Neumann but others were real bugs in the original programs. This demonstrates once again that unless published program code has actually been run on a digital computer, it is probably wrong. The new article contains further references, one of which is about the emulator itself, and is included below. There is also an important earlier third paper that is not cited by the new article. Here is a summary of those references, with links: Donald E. Knuth [John] von Neumann's first computer program ACM Computing Surveys 2(4) 247--260 December 1970 http://doi.acm.org/10.1145/356580.356581 Barry S. Fagin and Dale J. Skrien IASSim: a Programmable Emulator for the Princeton IAS/von Neumann Machine Proceedings of the 42nd ACM Technical Symposium on Computer Science Education (SIGCSE 11), 359--364 (2011) http://dx.doi.org/10.1145/1953163.1953271 http://www.cs.colby.edu/djskrien/CPUSim/ http://www.cs.colby.edu/djskrien/IASSim/ Barry Fagin and Dale Skrien Debugging on the Shoulders of Giants: von Neumann's Programs 65 Years Later [IEEE] Computer 45(11) 59--68 November 2012 http://doi.ieeecomputersociety.org/10.1109/MC.2012.69 ---------------------------------------- P.S. I maintain many bibliographies, one of which is on virtual machines: http://www.math.utah.edu/pub/tex/bib/virtual-machines.bib http://www.math.utah.edu/pub/tex/bib/virtual-machines.html [They look similar on the screen, but the HTML version has live hypertext links, and the .bib version is needed for BibTeX processing.] That bibliography was first created on 10-Apr-2006, and now has almost 1000 entries. Its contents are mined from an SQL database representing a collection of almost 700,000 entries covering several hundred journals, plus many author- and subject-specific bibliographies. There is also a complete bibliography of von Neumann's works in the BibNet Project archives: http://www.math.utah.edu/pub/bibnet/authors/v/von-neumann-john.bib http://www.math.utah.edu/pub/bibnet/authors/v/von-neumann-john.html ------------------------------------------------------------------------------- - Nelson H. F. Beebe Tel: +1 801 581 5254 - - University of Utah FAX: +1 801 581 4148 - - Department of Mathematics, 110 LCB Internet e-mail: beebe at math.utah.edu - - 155 S 1400 E RM 233 beebe at acm.org beebe at computer.org - - Salt Lake City, UT 84112-0090, USA URL: http://www.math.utah.edu/~beebe/ - ------------------------------------------------------------------------------- From lbickley at bickleywest.com Sat Nov 17 13:51:54 2012 From: lbickley at bickleywest.com (Lyle Bickley) Date: Sat, 17 Nov 2012 10:51:54 -0800 Subject: [Simh] new papers on von Neumann's programs and machine In-Reply-To: References: Message-ID: <20121117105154.2c779c49@asrock.bcwi.net> On Sat, 17 Nov 2012 09:15:47 -0700 (MST) "Nelson H. F. Beebe" wrote: > [This posting is slightly off-target, because it is about virtual > machines, and not Simh specifically. Nevertheless, the poster feels > that it may well be of interest to list readers.] > > The latest issue of the IEEE journal Computer arrived in my mailbox a > couple of days ago. It contains an interesting article that describes > an emulator for John von Neumann's IAS (Institute for Advanced Study) > computer, designed in the late 1940s, but not fully operational until > 1952 (see, e.g., > > http://en.wikipedia.org/wiki/IAS_machine > > ). > > The article also describes the testing of several early programs > written by von Neumann and his close collaborator, Herman Goldstine. > The article authors report finding several bugs in the programs (all > of which were published about five years before actual hardware became > operational). Some of the errors arise from transcription from > technical reports into the volumes of the Collected Works of John von > Neumann but others were real bugs in the original programs. This > demonstrates once again that unless published program code has > actually been run on a digital computer, it is probably wrong. > > The new article contains further references, one of which is about the > emulator itself, and is included below. There is also an important > earlier third paper that is not cited by the new article. Here is a > summary of those references, with links: > > Donald E. Knuth > [John] von Neumann's first computer program > ACM Computing Surveys 2(4) 247--260 December 1970 > http://doi.acm.org/10.1145/356580.356581 > > Barry S. Fagin and Dale J. Skrien > IASSim: a Programmable Emulator for the Princeton IAS/von > Neumann Machine Proceedings of the 42nd ACM Technical Symposium on > Computer Science Education (SIGCSE 11), 359--364 (2011) > http://dx.doi.org/10.1145/1953163.1953271 > http://www.cs.colby.edu/djskrien/CPUSim/ > http://www.cs.colby.edu/djskrien/IASSim/ Thanks for sharing these links! I've been playing with the IAS Simulator, and it is an excellent tool to become familiar with what it was like to program the IAS! I plan on passing this information on to the volunteers at the Computer History Museum! (I'm a member of the PDP-1 Restoration Team). I will, of course, give you credit for bringing it to my attention! --snip-- Cheers, Lyle Bickley Consulting West Inc. http://bickleywest.com "Black holes are where God is dividing by zero" From tivo.overo at gmail.com Sat Nov 17 15:14:24 2012 From: tivo.overo at gmail.com (Gary Lee Phillips) Date: Sat, 17 Nov 2012 14:14:24 -0600 Subject: [Simh] VAX XQ and wireless networking Message-ID: Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the VAX with OpenVMS 7.3. It works great, except for networking. The hardware is ASUS EeePC with Atheros AR5001 wireless network adapter and Atheros L2 Fast Ethernet adapter. I have libpcap version 1.1.1 (which is also callable as version 0.8) installed on Linux. The VAX module was compiled with networking enabled. The configuration file has an "attach xq0 wlan0" or "attach xq0 eth0" depending on which interface I'm trying to use. With the wired ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With the wireless (wlan0) neither one works. I've tried disabling the DECNET IV compatibility mode, which was recommended in one source for wireless compatibility, and also tried adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of the wireless adapter) to the configuration. The first did nothing I could discern, and the second generated an error message during startup of the SIMH module, saying that the MAC address was already in use and could not be assigned. Obviously the workaround is to use a cabled connection when the VAX emulator is needed, but this isn't very practical in most cases. Is there any other possibility I've overlooked? I notice that this is a frequently found question when using Google to look for answers, but I found no responses other than the two solutions mentioned above. I suppose a different network card emulation for VAX could make the wi-fi adapter look like a normal ethernet interface, but apparently that hasn't been done yet and it's probably beyond my very rudimentary C coding abilities. From dgahling at hotmail.com Sat Nov 17 15:24:02 2012 From: dgahling at hotmail.com (Dan Gahlinger) Date: Sat, 17 Nov 2012 15:24:02 -0500 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: Message-ID: It works...use the TUN interface and the bridging, works just fine. Dan. > Date: Sat, 17 Nov 2012 14:14:24 -0600 > From: tivo.overo at gmail.com > To: simh at trailing-edge.com > Subject: [Simh] VAX XQ and wireless networking > > Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the > VAX with OpenVMS 7.3. It works great, except for networking. The > hardware is ASUS EeePC with Atheros AR5001 wireless network adapter > and Atheros L2 Fast Ethernet adapter. > > I have libpcap version 1.1.1 (which is also callable as version 0.8) > installed on Linux. The VAX module was compiled with networking > enabled. > > The configuration file has an "attach xq0 wlan0" or "attach xq0 eth0" > depending on which interface I'm trying to use. With the wired > ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With the > wireless (wlan0) neither one works. > > I've tried disabling the DECNET IV compatibility mode, which was > recommended in one source for wireless compatibility, and also tried > adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of the > wireless adapter) to the configuration. The first did nothing I could > discern, and the second generated an error message during startup of > the SIMH module, saying that the MAC address was already in use and > could not be assigned. > > Obviously the workaround is to use a cabled connection when the VAX > emulator is needed, but this isn't very practical in most cases. Is > there any other possibility I've overlooked? > > I notice that this is a frequently found question when using Google to > look for answers, but I found no responses other than the two > solutions mentioned above. I suppose a different network card > emulation for VAX could make the wi-fi adapter look like a normal > ethernet interface, but apparently that hasn't been done yet and it's > probably beyond my very rudimentary C coding abilities. > _______________________________________________ > Simh mailing list > Simh at trailing-edge.com > http://mailman.trailing-edge.com/mailman/listinfo/simh -------------- next part -------------- An HTML attachment was scrubbed... URL: From bqt at softjar.se Sat Nov 17 15:57:46 2012 From: bqt at softjar.se (Johnny Billquist) Date: Sat, 17 Nov 2012 21:57:46 +0100 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: Message-ID: <50A7FA4A.9080607@softjar.se> On 2012-11-17 21:14, Gary Lee Phillips wrote: > Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the > VAX with OpenVMS 7.3. It works great, except for networking. The > hardware is ASUS EeePC with Atheros AR5001 wireless network adapter > and Atheros L2 Fast Ethernet adapter. > > I have libpcap version 1.1.1 (which is also callable as version 0.8) > installed on Linux. The VAX module was compiled with networking > enabled. > > The configuration file has an "attach xq0 wlan0" or "attach xq0 eth0" > depending on which interface I'm trying to use. With the wired > ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With the > wireless (wlan0) neither one works. > > I've tried disabling the DECNET IV compatibility mode, which was > recommended in one source for wireless compatibility, and also tried > adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of the > wireless adapter) to the configuration. The first did nothing I could > discern, and the second generated an error message during startup of > the SIMH module, saying that the MAC address was already in use and > could not be assigned. > > Obviously the workaround is to use a cabled connection when the VAX > emulator is needed, but this isn't very practical in most cases. Is > there any other possibility I've overlooked? > > I notice that this is a frequently found question when using Google to > look for answers, but I found no responses other than the two > solutions mentioned above. I suppose a different network card > emulation for VAX could make the wi-fi adapter look like a normal > ethernet interface, but apparently that hasn't been done yet and it's > probably beyond my very rudimentary C coding abilities. The "problem" is that a wlan interface is not really the same thing as an ethernet interface. As Dan suggested (I noticed), use a tun interface, and let than tunnel into the wlan instead, and things might work better. Johnny -- Johnny Billquist || "I'm on a bus || on a psychedelic trip email: bqt at softjar.se || Reading murder books pdp is alive! || tryin' to stay hip" - B. Idol From tivo.overo at gmail.com Sat Nov 17 15:58:13 2012 From: tivo.overo at gmail.com (Gary Lee Phillips) Date: Sat, 17 Nov 2012 14:58:13 -0600 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: Message-ID: I should have mentioned that I've tried setting up bridging and taptap without success when using the wireless interface. I use it on my desktop PCs (wired ethernet and Ubuntu) where it works OK though it's very cumbersome to start up and shut down. I thought the main advantage of a bridge was the fact that it lets the host and the emulated machine speak directly to one another via network packets, and that's a feature I don't really need in this case. Using the serial port emulation with a VT100 emulation over a tcp/ip socket is adequate and works well. On the desktop machines, the taptap module seems to have a tendency to "hang" on shutdown, even though the SIMH emulation has been closed down normally. The unreleased bridge port will keep the entire host from shutting down and requires a forced power off, which isn't very nice. On 11/17/12, Dan Gahlinger wrote: > > It works...use the TUN interface and the bridging, works just fine. > Dan. > >> Date: Sat, 17 Nov 2012 14:14:24 -0600 >> From: tivo.overo at gmail.com >> To: simh at trailing-edge.com >> Subject: [Simh] VAX XQ and wireless networking >> >> Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the >> VAX with OpenVMS 7.3. It works great, except for networking. The >> hardware is ASUS EeePC with Atheros AR5001 wireless network adapter >> and Atheros L2 Fast Ethernet adapter. >> >> I have libpcap version 1.1.1 (which is also callable as version 0.8) >> installed on Linux. The VAX module was compiled with networking >> enabled. >> >> The configuration file has an "attach xq0 wlan0" or "attach xq0 eth0" >> depending on which interface I'm trying to use. With the wired >> ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With the >> wireless (wlan0) neither one works. >> >> I've tried disabling the DECNET IV compatibility mode, which was >> recommended in one source for wireless compatibility, and also tried >> adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of the >> wireless adapter) to the configuration. The first did nothing I could >> discern, and the second generated an error message during startup of >> the SIMH module, saying that the MAC address was already in use and >> could not be assigned. >> >> Obviously the workaround is to use a cabled connection when the VAX >> emulator is needed, but this isn't very practical in most cases. Is >> there any other possibility I've overlooked? >> >> I notice that this is a frequently found question when using Google to >> look for answers, but I found no responses other than the two >> solutions mentioned above. I suppose a different network card >> emulation for VAX could make the wi-fi adapter look like a normal >> ethernet interface, but apparently that hasn't been done yet and it's >> probably beyond my very rudimentary C coding abilities. >> _______________________________________________ >> Simh mailing list >> Simh at trailing-edge.com >> http://mailman.trailing-edge.com/mailman/listinfo/simh > From dgahling at hotmail.com Sat Nov 17 16:03:56 2012 From: dgahling at hotmail.com (Dan Gahlinger) Date: Sat, 17 Nov 2012 16:03:56 -0500 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: , , Message-ID: actually that's exactly what you DO need to do. I do this myself with a wireless wlan interface and it works just fine. you need the bridge and attach to the TUN/TAP interface.and as the other post mentioned bridge the wlan interface the unreleased bridge port certainly will not keep the entire host from shutting down,I have no idea where you got such a crazy idea. it works just fine for the two systems I run this on wireless. I think this could even be made to work niceless on a raspberry pi rig. imagine that a vax-cluster of wireless embedded systems, very cool. > Date: Sat, 17 Nov 2012 14:58:13 -0600 > Subject: Re: [Simh] VAX XQ and wireless networking > From: tivo.overo at gmail.com > To: dgahling at hotmail.com > CC: simh at trailing-edge.com > > I should have mentioned that I've tried setting up bridging and taptap > without success when using the wireless interface. I use it on my > desktop PCs (wired ethernet and Ubuntu) where it works OK though it's > very cumbersome to start up and shut down. > > I thought the main advantage of a bridge was the fact that it lets the > host and the emulated machine speak directly to one another via > network packets, and that's a feature I don't really need in this > case. Using the serial port emulation with a VT100 emulation over a > tcp/ip socket is adequate and works well. > > On the desktop machines, the taptap module seems to have a tendency to > "hang" on shutdown, even though the SIMH emulation has been closed > down normally. The unreleased bridge port will keep the entire host > from shutting down and requires a forced power off, which isn't very > nice. > > On 11/17/12, Dan Gahlinger wrote: > > > > It works...use the TUN interface and the bridging, works just fine. > > Dan. > > > >> Date: Sat, 17 Nov 2012 14:14:24 -0600 > >> From: tivo.overo at gmail.com > >> To: simh at trailing-edge.com > >> Subject: [Simh] VAX XQ and wireless networking > >> > >> Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the > >> VAX with OpenVMS 7.3. It works great, except for networking. The > >> hardware is ASUS EeePC with Atheros AR5001 wireless network adapter > >> and Atheros L2 Fast Ethernet adapter. > >> > >> I have libpcap version 1.1.1 (which is also callable as version 0.8) > >> installed on Linux. The VAX module was compiled with networking > >> enabled. > >> > >> The configuration file has an "attach xq0 wlan0" or "attach xq0 eth0" > >> depending on which interface I'm trying to use. With the wired > >> ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With the > >> wireless (wlan0) neither one works. > >> > >> I've tried disabling the DECNET IV compatibility mode, which was > >> recommended in one source for wireless compatibility, and also tried > >> adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of the > >> wireless adapter) to the configuration. The first did nothing I could > >> discern, and the second generated an error message during startup of > >> the SIMH module, saying that the MAC address was already in use and > >> could not be assigned. > >> > >> Obviously the workaround is to use a cabled connection when the VAX > >> emulator is needed, but this isn't very practical in most cases. Is > >> there any other possibility I've overlooked? > >> > >> I notice that this is a frequently found question when using Google to > >> look for answers, but I found no responses other than the two > >> solutions mentioned above. I suppose a different network card > >> emulation for VAX could make the wi-fi adapter look like a normal > >> ethernet interface, but apparently that hasn't been done yet and it's > >> probably beyond my very rudimentary C coding abilities. > >> _______________________________________________ > >> Simh mailing list > >> Simh at trailing-edge.com > >> http://mailman.trailing-edge.com/mailman/listinfo/simh > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From oboguev at yahoo.com Sat Nov 17 16:04:42 2012 From: oboguev at yahoo.com (Sergey Oboguev) Date: Sat, 17 Nov 2012 13:04:42 -0800 (PST) Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: Message-ID: <1353186282.39615.YahooMailRC@web184301.mail.ne1.yahoo.com> I have not personally tried it, but here are few general remarks: SIMH DEQNA/DELQA virtual devices talk 802.3, whereas WiFi (802.11) is different from 802.3 in frame structure and everything else. WiFi may superficially look like Ethernet (i.e. have similarly looking MAC address), but it is actually not, so naive binding of XQ to host WiFi interface would not work for that reason. In addition, guest operating system software using DEQNA/DELQA will likewise be expecting 802.3 structured frames and protocol capabilities. Therefore, some form of translation between 802.3 and 802.11 or tunneling of 802.3 packets through 802.11 (or generic TCP/IP) would be required. If all you want is to connect SIMH instance to exterior TCP/IP network via IP connectivity only (i.e. do not need VAXcluster, LAT or DECnet), then regular host OS level bridging of tap interface (or Windows interface such as loopback or wired Ethernet interfaces) to WiFi interface would be one thing to try. Bridging is likely to suffice to translate IP packets from 802.3 MAC envelope to 802.11 and vice versa; that is it's purpose anyway. If, on the other hand, you want to interconnect multiple SIMH instances over WiFi utilizing not only IP, but also other Ethernet based protocols such as VAXcluster or LAT or DECnet, host OS bridging may fail to translate those protocols. In the latter case you'd need to find a way to tunnel 802.3 packets over 802.11 or over generic TCP/IP. General approach would be to attach SIMH XQ to Linux tap device on one side and hook tap/tun's other side into tunneling software. VDE (http://wiki.virtualsquare.org/wiki/index.php/VDE) does just that, but I personally do not have experience using VDE. Furthermore, there is some direct VDE support in the latest version of SIMH (assuming the simulator executable is built with VDE option enabled). I have not personally tried using it, but at a glance at the code, it looks like it allows to connect XQ to VDE_SWITCH directly (using vde:port syntax), not requiring tap device as an intermediary. -------------- next part -------------- An HTML attachment was scrubbed... URL: From tivo.overo at gmail.com Sat Nov 17 16:31:50 2012 From: tivo.overo at gmail.com (Gary Lee Phillips) Date: Sat, 17 Nov 2012 15:31:50 -0600 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: Message-ID: Might be some peculiarity of Ubuntu or my two Dell desktop PCs, but the system logs do in fact show that they are waiting for the tun0 device to close down and that it is "busy". I've waited as long as 30 minutes after the SIMH module was closed, and still no shutdown for Linux. The desktop machines have SIMH 3.8 rather than 3.9, which might also make some difference. I will try the bridge again, but previous attempts on the laptop failed to work for either the host or the guest system, using the same sequence of steps to set up the bridge that work with the desktop machines and wired ethernet. On 11/17/12, Dan Gahlinger wrote: > > actually that's exactly what you DO need to do. > I do this myself with a wireless wlan interface and it works just fine. > you need the bridge and attach to the TUN/TAP interface.and as the other > post mentioned bridge the wlan interface > the unreleased bridge port certainly will not keep the entire host from > shutting down,I have no idea where you got such a crazy idea. > it works just fine for the two systems I run this on wireless. > I think this could even be made to work niceless on a raspberry pi rig. > imagine that a vax-cluster of wireless embedded systems, very cool. From jg at jordi.guillaumes.name Sat Nov 17 16:46:40 2012 From: jg at jordi.guillaumes.name (Jordi Guillaumes i Pons) Date: Sat, 17 Nov 2012 22:46:40 +0100 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: <1353186282.39615.YahooMailRC@web184301.mail.ne1.yahoo.com> References: <1353186282.39615.YahooMailRC@web184301.mail.ne1.yahoo.com> Message-ID: <50A805C0.7000209@jordi.guillaumes.name> Al 17/11/12 22:04, En/na Sergey Oboguev ha escrit: > In the latter case you'd need to find a way to tunnel 802.3 packets > over 802.11 or over generic TCP/IP. General approach would be to > attach SIMH XQ to Linux tap device on one side and hook tap/tun's > other side into tunneling software. VDE > (http://wiki.virtualsquare.org/wiki/index.php/VDE) does just that, but > I personally do not have experience using VDE. Furthermore, there is > some direct VDE support in the latest version of SIMH (assuming the > simulator executable is built with VDE option enabled). I have not > personally tried using it, but at a glance at the code, it looks like > it allows to connect XQ to VDE_SWITCH directly (using vde:port > syntax), not requiring tap device as an intermediary. It does, and it works, but there is a caveat. I have not been successful bridging the tapx device which the vde switch exposes to the wlan physical interface. It is easily done with a real, wired ethernet, but not with a wlan. There is a solution. VDE comes with some support utilities which allow to "plug" into the virtual switch. Using vde_plug and dpipe (another VDE utility which provides two-way redirection between two unix processes) you can establish a ssh tunnel which acts as a transport pipe for the raw ethernet packages used by VDE. It works at level 2, so it should provide DECNET, LAT and LAVC connectivity (I have not tried the last one). As a bonus, it works with any IP link, including a tethered 3D link or even a POTS line. From simh at tj.merritts.org Sat Nov 17 18:07:06 2012 From: simh at tj.merritts.org (TJ Merritt) Date: Sat, 17 Nov 2012 15:07:06 -0800 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: References: Message-ID: <1353193626.2301.3.camel@vb-tjm> The TUN interface will get the packets injected into the system, but will not get them bridged across the wireless network which uses the MAC address for managing the wireless network. Since your MAC address for the virtual interface is not the same as the mac address for the host, reply packets over wireless will get dropped. The trick is to get the packets to go out with the host's MAC address and replies routed back to the virtual interface. VirtualBox does this by sending packets out with the original IP address but the MAC address of the host, they then record the virtual interface that the IP address came from, when the reply packet comes back, they use the IP address to look up the virtual interface associated with that IP address and route the packet to that interface. There are special cases of course for things like broadcast, multicast, and IP addresses that are shared between multiple VM's. VirtualBox does not currently support IPv6 addresses though. This would also not help for DECnet. This is effectively a layer 3 switching capability that works around the lack of layer 2 switching in the wireless network. Another approach would be to setup NAT so that the NAT layer can take care of the MAC address rewriting. It will also rewrite the IP address though. If this is satisfactory, you there should be plenty of NAT resources available to you on Mint such that no code writing is required. Adding virutal networking support for wireless networks would be a major undertaking for SimH, but if done, could eliminate most of the hiccups with virtual networking over wireless networks. Of course, not of these issues occur over wired netowrks. -- TJ On Sat, 2012-11-17 at 15:24 -0500, Dan Gahlinger wrote: > It works... > use the TUN interface and the bridging, works just fine. > > > Dan. > > > > Date: Sat, 17 Nov 2012 14:14:24 -0600 > > From: tivo.overo at gmail.com > > To: simh at trailing-edge.com > > Subject: [Simh] VAX XQ and wireless networking > > > > Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the > > VAX with OpenVMS 7.3. It works great, except for networking. The > > hardware is ASUS EeePC with Atheros AR5001 wireless network adapter > > and Atheros L2 Fast Ethernet adapter. > > > > I have libpcap version 1.1.1 (which is also callable as version 0.8) > > installed on Linux. The VAX module was compiled with networking > > enabled. > > > > The configuration file has an "attach xq0 wlan0" or "attach xq0 > eth0" > > depending on which interface I'm trying to use. With the wired > > ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With > the > > wireless (wlan0) neither one works. > > > > I've tried disabling the DECNET IV compatibility mode, which was > > recommended in one source for wireless compatibility, and also tried > > adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of > the > > wireless adapter) to the configuration. The first did nothing I > could > > discern, and the second generated an error message during startup of > > the SIMH module, saying that the MAC address was already in use and > > could not be assigned. > > > > Obviously the workaround is to use a cabled connection when the VAX > > emulator is needed, but this isn't very practical in most cases. Is > > there any other possibility I've overlooked? > > > > I notice that this is a frequently found question when using Google > to > > look for answers, but I found no responses other than the two > > solutions mentioned above. I suppose a different network card > > emulation for VAX could make the wi-fi adapter look like a normal > > ethernet interface, but apparently that hasn't been done yet and > it's > > probably beyond my very rudimentary C coding abilities. > > _______________________________________________ > > Simh mailing list > > Simh at trailing-edge.com > > http://mailman.trailing-edge.com/mailman/listinfo/simh > > _______________________________________________ > Simh mailing list > Simh at trailing-edge.com > http://mailman.trailing-edge.com/mailman/listinfo/simh From bqt at softjar.se Sat Nov 17 18:54:10 2012 From: bqt at softjar.se (Johnny Billquist) Date: Sun, 18 Nov 2012 00:54:10 +0100 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: <1353193626.2301.3.camel@vb-tjm> References: <1353193626.2301.3.camel@vb-tjm> Message-ID: <50A823A2.1050602@softjar.se> If that's how VirtualBox works, then it is rather broken. How about when you talk DECnet? There are no IP addresses involved at all. Johnny On 2012-11-18 00:07, TJ Merritt wrote: > The TUN interface will get the packets injected into the system, but > will not get them bridged across the wireless network which uses the MAC > address for managing the wireless network. Since your MAC address for > the virtual interface is not the same as the mac address for the host, > reply packets over wireless will get dropped. The trick is to get the > packets to go out with the host's MAC address and replies routed back to > the virtual interface. > > VirtualBox does this by sending packets out with the original IP address > but the MAC address of the host, they then record the virtual interface > that the IP address came from, when the reply packet comes back, they > use the IP address to look up the virtual interface associated with that > IP address and route the packet to that interface. There are special > cases of course for things like broadcast, multicast, and IP addresses > that are shared between multiple VM's. VirtualBox does not currently > support IPv6 addresses though. This would also not help for DECnet. > This is effectively a layer 3 switching capability that works around the > lack of layer 2 switching in the wireless network. > > Another approach would be to setup NAT so that the NAT layer can take > care of the MAC address rewriting. It will also rewrite the IP address > though. If this is satisfactory, you there should be plenty of NAT > resources available to you on Mint such that no code writing is > required. > > Adding virutal networking support for wireless networks would be a major > undertaking for SimH, but if done, could eliminate most of the hiccups > with virtual networking over wireless networks. > > Of course, not of these issues occur over wired netowrks. > > -- TJ > > On Sat, 2012-11-17 at 15:24 -0500, Dan Gahlinger wrote: >> It works... >> use the TUN interface and the bridging, works just fine. >> >> >> Dan. >> >> >>> Date: Sat, 17 Nov 2012 14:14:24 -0600 >>> From: tivo.overo at gmail.com >>> To: simh at trailing-edge.com >>> Subject: [Simh] VAX XQ and wireless networking >>> >>> Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the >>> VAX with OpenVMS 7.3. It works great, except for networking. The >>> hardware is ASUS EeePC with Atheros AR5001 wireless network adapter >>> and Atheros L2 Fast Ethernet adapter. >>> >>> I have libpcap version 1.1.1 (which is also callable as version 0.8) >>> installed on Linux. The VAX module was compiled with networking >>> enabled. >>> >>> The configuration file has an "attach xq0 wlan0" or "attach xq0 >> eth0" >>> depending on which interface I'm trying to use. With the wired >>> ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With >> the >>> wireless (wlan0) neither one works. >>> >>> I've tried disabling the DECNET IV compatibility mode, which was >>> recommended in one source for wireless compatibility, and also tried >>> adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of >> the >>> wireless adapter) to the configuration. The first did nothing I >> could >>> discern, and the second generated an error message during startup of >>> the SIMH module, saying that the MAC address was already in use and >>> could not be assigned. >>> >>> Obviously the workaround is to use a cabled connection when the VAX >>> emulator is needed, but this isn't very practical in most cases. Is >>> there any other possibility I've overlooked? >>> >>> I notice that this is a frequently found question when using Google >> to >>> look for answers, but I found no responses other than the two >>> solutions mentioned above. I suppose a different network card >>> emulation for VAX could make the wi-fi adapter look like a normal >>> ethernet interface, but apparently that hasn't been done yet and >> it's >>> probably beyond my very rudimentary C coding abilities. >>> _______________________________________________ >>> Simh mailing list >>> Simh at trailing-edge.com >>> http://mailman.trailing-edge.com/mailman/listinfo/simh >> >> _______________________________________________ >> Simh mailing list >> Simh at trailing-edge.com >> http://mailman.trailing-edge.com/mailman/listinfo/simh > > > _______________________________________________ > Simh mailing list > Simh at trailing-edge.com > http://mailman.trailing-edge.com/mailman/listinfo/simh > -- Johnny Billquist || "I'm on a bus || on a psychedelic trip email: bqt at softjar.se || Reading murder books pdp is alive! || tryin' to stay hip" - B. Idol From tivo.overo at gmail.com Sat Nov 17 19:02:48 2012 From: tivo.overo at gmail.com (Gary Lee Phillips) Date: Sat, 17 Nov 2012 18:02:48 -0600 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: <1353193626.2301.3.camel@vb-tjm> References: <1353193626.2301.3.camel@vb-tjm> Message-ID: Well, just to confirm what I thought I remembered: I tried again to set up bridging, and it is not possible using the method that works for wired ethernet. That is, I define a bridge br0 using the brctl command (as root of course) and then add interfaces to the bridge. I can add eth0 without difficulty, just as on other machines. But if I try to add the wireless interface wlan0, the command is rejected: can't add wlan0 to bridge br0: Operation not supported I have used iptables (and earlier versions of the command, ipchains or ifwadm) to set up software firewalls before, and even to NAT ip addresses, but I've never seen a way to make it change a MAC address. Even if I had, I have no idea how to apply that to the output of the SIMH XQ device. Are you saying I should create a tun device or something similar to which XQ0 can be attached, and then re-route packets from that device to wlan0 by using iptables? I believe if IP were working, then DECNET can be routed or encapsulated over IP. But in this particular case, DECNET is inessential if I have working IP. The working IP tools and clients are sufficient. And I thought DECNET Plus didn't need to mess with MAC addresses anyway. Isn't it just DECNET IV that does that? On 11/17/12, TJ Merritt wrote: > The TUN interface will get the packets injected into the system, but > will not get them bridged across the wireless network which uses the MAC > address for managing the wireless network. Since your MAC address for > the virtual interface is not the same as the mac address for the host, > reply packets over wireless will get dropped. The trick is to get the > packets to go out with the host's MAC address and replies routed back to > the virtual interface. > > VirtualBox does this by sending packets out with the original IP address > but the MAC address of the host, they then record the virtual interface > that the IP address came from, when the reply packet comes back, they > use the IP address to look up the virtual interface associated with that > IP address and route the packet to that interface. There are special > cases of course for things like broadcast, multicast, and IP addresses > that are shared between multiple VM's. VirtualBox does not currently > support IPv6 addresses though. This would also not help for DECnet. > This is effectively a layer 3 switching capability that works around the > lack of layer 2 switching in the wireless network. > > Another approach would be to setup NAT so that the NAT layer can take > care of the MAC address rewriting. It will also rewrite the IP address > though. If this is satisfactory, you there should be plenty of NAT > resources available to you on Mint such that no code writing is > required. > > Adding virutal networking support for wireless networks would be a major > undertaking for SimH, but if done, could eliminate most of the hiccups > with virtual networking over wireless networks. > > Of course, not of these issues occur over wired netowrks. > > -- TJ > > On Sat, 2012-11-17 at 15:24 -0500, Dan Gahlinger wrote: >> It works... >> use the TUN interface and the bridging, works just fine. >> >> >> Dan. >> >> >> > Date: Sat, 17 Nov 2012 14:14:24 -0600 >> > From: tivo.overo at gmail.com >> > To: simh at trailing-edge.com >> > Subject: [Simh] VAX XQ and wireless networking >> > >> > Is this even possible? I have SIMH 3.9 on Linux Mint 12, running the >> > VAX with OpenVMS 7.3. It works great, except for networking. The >> > hardware is ASUS EeePC with Atheros AR5001 wireless network adapter >> > and Atheros L2 Fast Ethernet adapter. >> > >> > I have libpcap version 1.1.1 (which is also callable as version 0.8) >> > installed on Linux. The VAX module was compiled with networking >> > enabled. >> > >> > The configuration file has an "attach xq0 wlan0" or "attach xq0 >> eth0" >> > depending on which interface I'm trying to use. With the wired >> > ethernet (eth0) I am able to use both TCP/IP and DECNET Plus. With >> the >> > wireless (wlan0) neither one works. >> > >> > I've tried disabling the DECNET IV compatibility mode, which was >> > recommended in one source for wireless compatibility, and also tried >> > adding "set xq xx-yy-zz-aa-bb-cc" (with the actual MAC address of >> the >> > wireless adapter) to the configuration. The first did nothing I >> could >> > discern, and the second generated an error message during startup of >> > the SIMH module, saying that the MAC address was already in use and >> > could not be assigned. >> > >> > Obviously the workaround is to use a cabled connection when the VAX >> > emulator is needed, but this isn't very practical in most cases. Is >> > there any other possibility I've overlooked? >> > >> > I notice that this is a frequently found question when using Google >> to >> > look for answers, but I found no responses other than the two >> > solutions mentioned above. I suppose a different network card >> > emulation for VAX could make the wi-fi adapter look like a normal >> > ethernet interface, but apparently that hasn't been done yet and >> it's >> > probably beyond my very rudimentary C coding abilities. >> > _______________________________________________ >> > Simh mailing list >> > Simh at trailing-edge.com >> > http://mailman.trailing-edge.com/mailman/listinfo/simh >> >> _______________________________________________ >> Simh mailing list >> Simh at trailing-edge.com >> http://mailman.trailing-edge.com/mailman/listinfo/simh > > > From simh at tj.merritts.org Sun Nov 18 10:10:21 2012 From: simh at tj.merritts.org (TJ Merritt) Date: Sun, 18 Nov 2012 07:10:21 -0800 Subject: [Simh] VAX XQ and wireless networking In-Reply-To: <50A823A2.1050602@softjar.se> References: <1353193626.2301.3.camel@vb-tjm> <50A823A2.1050602@softjar.se> Message-ID: <1353251421.2301.6.camel@vb-tjm> It's not so much that VirtualBox is broken, but rather it is working around the brokenness in 802.11. VirtualBox has hooks for supporting L3 bridging of any L3 protocol, but there is only an implementation for IPv4. I'm aware of this issue only because of the annoyance of IPv6 bridging failing on wireless on VirtualBox VM's where IPv4 works. Overall the networking support in Virtual Box is quite superior. -- TJ On Sun, 2012-11-18 at 00:54 +0100, Johnny Billquist wrote: > If that's how VirtualBox works, then it is rather broken. > How about when you talk DECnet? There are no IP addresses involved at all. > > Johnny > From jg at jordi.guillaumes.name Mon Nov 19 12:16:17 2012 From: jg at jordi.guillaumes.name (Jordi Guillaumes i Pons) Date: Mon, 19 Nov 2012 18:16:17 +0100 Subject: [Simh] Possible problem in Extra-VAXen branch: Message-ID: <50AA6961.6090404@jordi.guillaumes.name> vax780 seems to be broken since a few days ago: jguillaumes at pi01 ~/emul/vax780 $ ../simh/BIN/vax780 vax780.ini VAX780 simulator V4.0-0 libpcap version 1.3.0 Eth: opened OS device vde:/tmp/vde.ctl Listening on port 32123 (socket 8) Modem control activated Auto disconnect activated sim> b rp Loading boot code from vmb.exe Saving boot code to vmb.exe Loading boot code from vmb.exe Address space exceeded, PC: 00000000 (HALT) Backing to the Oct 31 commit (69666f1480d494c225a1e82b8b38568792948382): jguillaumes at pi01 ~/emul/vax780 $ ../simh/BIN/vax780 vax780.ini VAX780 simulator V4.0-0 libpcap version 1.3.0 Eth: opened OS device vde:/tmp/vde.ctl Listening on port 32123 (socket 8) Modem control activated Auto disconnect activated sim> b rp Loading boot code from vmb.exe VAX/VMS Version V4.7 28-Oct-1987 13:00 -------------- next part -------------- An HTML attachment was scrubbed... URL: From petermallan at gmail.com Wed Nov 21 15:27:27 2012 From: petermallan at gmail.com (Peter Allan) Date: Wed, 21 Nov 2012 20:27:27 +0000 Subject: [Simh] Adding a third massbus to the VAX780 emulator Message-ID: Many, many years ago, I used to be the system manager of a VAX-11/780 system that was part of the Starlink astronomical data processing network in the UK. One of my pet projects is to have an emulation of this as close to the original as I can get it. The system had two RP06 disk drives and two System Industries drives that emulated RM05 drives. Since the System Industries drives interfaced to the 780 through their own massbus adapter, the RP06 drives appeared as devices DBA0 and DBA1, and the emulated RM05 drives appeared as DRA0 and DRA1. I can create a similar system using simh, but with all of the drives on a single massbus, the drives appear as devices DBA0, DBA1, DRA2 and DRA3. This apparently trivial difference annoys me so I am looking for a way of having both the RP06 and RM05 drives appear as units 0 and 1. I presume that this will require modifying simh to add a third massbus adapter. (Please do tell me if there is an easier way.) Having looked at the simh code, I am reasonably confident of being able to add a third massbus. However, how can I attach RP drives to one massbus and RM drives to another? Both drives are handled by the same function, i.e. pdp11_rp.c. My thoughts at the moment are that I should copy pdp11_rp.c to a new file pdp11_rm.c and remove the code that handles the RM disk from pdp11_rp.c and vice versa for the RP code in pdp11_rm.c I can see that there is more work to do in terms of splitting the handing of interrupts and the like, but before diving in, does my basic strategy seem correct? It would be more like real life (I think) to be able to add RM and RP drives to both massbuses that handle disks, but I don't see how to do that without a lot more work. And before anyone says it, I know I am mad to be trying to do this. The effort is clearly disproportionate to the benefit, but it is a hobby after all. All comments gratefully received. Peter Allan, petermallan at gmail.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From bqt at softjar.se Wed Nov 21 16:03:17 2012 From: bqt at softjar.se (Johnny Billquist) Date: Wed, 21 Nov 2012 22:03:17 +0100 Subject: [Simh] Adding a third massbus to the VAX780 emulator In-Reply-To: References: Message-ID: <50AD4195.5000401@softjar.se> I think this is totally the wrong strategy. If anything, you should try to model this as close to how the actual hardware worked, and not do some special thing that happens to just match your very specific configuration. What you need is to split the massbus handling, as such, from the specific disks. You should then be able to create massbusses for you machine, and then you attach disks to the massbus. It is a hierarchy, and the config should reflect that. I think simh's current approach of presenting things in a very flat way is a bit too simplistic. I got annoyed the other day when I wanted more than one DZ11 on a configuration, and realized I could not configure the controllers as such at all, but instead configured the total number of serial lines, with an implicit overflow from one controller to the next. In real life, you might have a machine with four serial lines, but all attached to separate DZ11. Just specifying the endpoints without the ability to specify the configuration in the middle layers will cause issues like the ones you have, because simh then makes assumptions for you. (Such as having just one massbus, and all disks attached to that one.) In fact, I wonder if this will actually totally break something like RSX-11M, which do not allow mixed device massbuses. Ie. in 11M you cannot have both RM and RP disks on the same massbus. M+ allows this, but then, M+ is rather more advanced... :-) Johnny On 2012-11-21 21:27, Peter Allan wrote: > Many, many years ago, I used to be the system manager of a VAX-11/780 > system that was part of the Starlink astronomical data processing > network in the UK. One of my pet projects is to have an emulation of > this as close to the original as I can get it. The system had two RP06 > disk drives and two System Industries drives that emulated RM05 drives. > Since the System Industries drives interfaced to the 780 through their > own massbus adapter, the RP06 drives appeared as devices DBA0 and DBA1, > and the emulated RM05 drives appeared as DRA0 and DRA1. > I can create a similar system using simh, but with all of the drives on > a single massbus, the drives appear as devices DBA0, DBA1, DRA2 and > DRA3. This apparently trivial difference annoys me so I am looking for a > way of having both the RP06 and RM05 drives appear as units 0 and 1. > I presume that this will require modifying simh to add a third massbus > adapter. (Please do tell me if there is an easier way.) > Having looked at the simh code, I am reasonably confident of being able > to add a third massbus. However, how can I attach RP drives to one > massbus and RM drives to another? Both drives are handled by the same > function, i.e. pdp11_rp.c. My thoughts at the moment are that I should > copy pdp11_rp.c to a new file pdp11_rm.c and remove the code that > handles the RM disk from pdp11_rp.c and vice versa for the RP code in > pdp11_rm.c > I can see that there is more work to do in terms of splitting the > handing of interrupts and the like, but before diving in, does my basic > strategy seem correct? It would be more like real life (I think) to be > able to add RM and RP drives to both massbuses that handle disks, but I > don't see how to do that without a lot more work. > And before anyone says it, I know I am mad to be trying to do this. The > effort is clearly disproportionate to the benefit, but it is a hobby > after all. > All comments gratefully received. > Peter Allan, petermallan at gmail.com > > > _______________________________________________ > Simh mailing list > Simh at trailing-edge.com > http://mailman.trailing-edge.com/mailman/listinfo/simh > From petermallan at gmail.com Thu Nov 22 04:36:18 2012 From: petermallan at gmail.com (Peter Allan) Date: Thu, 22 Nov 2012 09:36:18 +0000 Subject: [Simh] Adding a third massbus to the VAX780 emulator In-Reply-To: <0CC6789C1C831B4C8CCFF49D45D7010F840266F4B1@REDROOF2.alohasunset.com> References: <0CC6789C1C831B4C8CCFF49D45D7010F840266F4B1@REDROOF2.alohasunset.com> Message-ID: Thanks to all for your comments. Kevin, yes I could define logical names that would let me refer to the drives using the names DRA0 and DRA1, but they would still show up as DRA2 and DRA3 on the output of "sho dev d" for example. I would like to do better if I can. Johnny, I agree in principle, but I really don't want to get into rewriting too much of SIMH. Mark, that sounds like a very useful suggestion and I will look into how the RQ controllers are implemented. Hopefully it will indeed let me add any drive to any massbus, which is as it should be. If I succeed, it should be generally useful. My original suggestion was clearly a "dirty hack" and even if it worked, would not be suitable for airing in public. Cheers Peter Allan On 21 November 2012 22:09, Mark Pizzolato - Info Comm wrote: > Hi Peter,**** > > ** ** > > I have no specific knowledge of how multiple mass busses may be ‘wired’ > into simh. However, once you figure that detail out. I suggest that you > consider the idea of extending the existing pdp11_rp.c to support multiple > ‘device’ instances where each device is attached to a desired massbus and > has its own set of units. You may want to look at how pdp11_rq.c supports > multiple RQ instances as an example. Meanwhile, if you’d ever like to have > your changes make it back into the general simh code base, then please > start from the latest pdp11_rp.c which can be found at: > https://github.com/simh/simh/blob/Extra-VAXen/PDP11/pdp11_rp.c and a zip > of the current Extra-VAXen branch at: > https://github.com/simh/simh/archive/Extra-VAXen.zip**** > > I am in the middle of tracking down a bug in this code right now ( > https://github.com/simh/simh/issues/14), so feel free to start looking at > things now, but be prepared to take an update to pdp11_rp.c when I resolve > the issue.**** > > ** ** > > Good Luck,**** > > ** ** > > **- **Mark Pizzolato**** > > ** ** > > *From:* simh-bounces at trailing-edge.com [mailto: > simh-bounces at trailing-edge.com] *On Behalf Of *Peter Allan > *Sent:* Wednesday, November 21, 2012 12:27 PM > *To:* simh at trailing-edge.com > *Subject:* [Simh] Adding a third massbus to the VAX780 emulator**** > > ** ** > > Many, many years ago, I used to be the system manager of a VAX-11/780 > system that was part of the Starlink astronomical data processing network > in the UK. One of my pet projects is to have an emet defulation of this > as close to the original as I can get it. The system had two RP06 disk > drives and two System Industries drives that emulated RM05 drives. Since > the System Industries drives interfaced to the 780 through their own > massbus adapter, the RP06 drives appeared as devices DBA0 and DBA1, and the > emulated RM05 drives appeared as DRA0 and DRA1.**** > > **** > > I can create a similar system using simh, but with all of the drives on a > single massbus, the drives appear as devices DBA0, DBA1, DRA2 and DRA3. > This apparently trivial difference annoys me so I am looking for a way of > having both the RP06 and RM05 drives appear as units 0 and 1.**** > > **** > > I presume that this will require modifying simh to add a third massbus > adapter. (Please do tell me if there is an easier way.)**** > > **** > > Having looked at the simh code, I am reasonably confident of being able to > add a third massbus. However, how can I attach RP drives to one massbus and > RM drives to another? Both drives are handled by the same function, i.e. > pdp11_rp.c. My thoughts at the moment are that I should copy pdp11_rp.c to > a new file pdp11_rm.c and remove the code that handles the RM disk from > pdp11_rp.c and vice versa for the RP code in pdp11_rm.c**** > > **** > > I can see that there is more work to do in terms of splitting the handing > of interrupts and the like, but before diving in, does my basic strategy > seem correct? It would be more like real life (I think) to be able to add > RM and RP drives to both massbuses that handle disks, but I don't see how > to do that without a lot more work.**** > > **** > > And before anyone says it, I know I am mad to be trying to do this. The > effort is clearly disproportionate to the benefit, but it is a hobby after > all.**** > > **** > > All comments gratefully received.**** > > **** > > Peter Allan, petermallan at gmail.com**** > -------------- next part -------------- An HTML attachment was scrubbed... URL: From ddyer-simh at real-me.net Mon Nov 26 13:50:51 2012 From: ddyer-simh at real-me.net (Dave Dyer) Date: Mon, 26 Nov 2012 10:50:51 -0800 Subject: [Simh] even older PDP-10 macro Message-ID: <20121126185059.4DA491E0100@mini-me.trailing-edge.com> For my particular PDP-10 archeology, I need an even older version of Macro-10 than I can find in the various archive tapes. My recollection is that I need version 47, which would have been current about 1975. Either a binary or source would probably do. I hope someone has an image of a system tape from that era that was thought to be not important because there were newer tapes available). (back story; Either macro 50+ had new bugs or Dec deliberately changed the semantics of macros. In either case, a compile that had complicated macros no longer worked on macro version 50+, and even in the day I had to use "r old:macro" to compile.) From tshoppa at wmata.com Mon Nov 26 14:53:44 2012 From: tshoppa at wmata.com (Shoppa, Tim) Date: Mon, 26 Nov 2012 19:53:44 +0000 Subject: [Simh] even older PDP-10 macro In-Reply-To: <20121126185059.4DA491E0100@mini-me.trailing-edge.com> References: <20121126185059.4DA491E0100@mini-me.trailing-edge.com> Message-ID: <303A17BD5F8FA34DA45EEC245271AC0B2524B9C3@JGEX2K10MBX2.wmata.local> One big syntax change that happened between Macro V52 and Macro V53, was that support for "line starting with comma is a comment" was removed. The change that removed this support, seems to be reversible, i.e. a later Macro version could have this support added back in. I don't know when the change for semantics of complex macros was changed. Most of my efforts with early Macro sources, are related to compiling the PDP-6 monitor sources that are available, and it's been a long time since I've messed with that. Can't help with V47. But my notes on Macro versions from 50A onwards, including where they are available in the archives: 50A(441) 1-Jul-1976 As a Non-DEC update on DECUS 10-301 to add a user pushdown stack Patch between DEC and Non-DEC seems reversible. 53(1020) 27-Feb-1978 Has edit history from 50A through 53. TOPS-20 release 3 53A(1152) 19-Jul-1979 TOPS-20 release 3A 53B(1242) 3-Aug-1983 TOPS-20 release 6.1 53B(1244) 28-Jan-1986 March 86 TOPS-10 CUSP tapes 53B(1244) 13-Nov-85 TOPS-20 KS UPD 53B(1247) 29-Aug-86 Sept 88 TOPS-10 CUSP tapes 53B(1252) 24-Jun-88 TOPS-10 7.04 TSU01 -----Original Message----- From: simh-bounces at trailing-edge.com [mailto:simh-bounces at trailing-edge.com] On Behalf Of Dave Dyer Sent: Monday, November 26, 2012 1:51 PM To: Simh at trailing-edge.com Subject: [Simh] even older PDP-10 macro For my particular PDP-10 archeology, I need an even older version of Macro-10 than I can find in the various archive tapes. My recollection is that I need version 47, which would have been current about 1975. Either a binary or source would probably do. I hope someone has an image of a system tape from that era that was thought to be not important because there were newer tapes available). (back story; Either macro 50+ had new bugs or Dec deliberately changed the semantics of macros. In either case, a compile that had complicated macros no longer worked on macro version 50+, and even in the day I had to use "r old:macro" to compile.) _______________________________________________ Simh mailing list Simh at trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh From ddyer-simh at real-me.net Mon Nov 26 15:22:24 2012 From: ddyer-simh at real-me.net (Dave Dyer) Date: Mon, 26 Nov 2012 12:22:24 -0800 Subject: [Simh] even older PDP-10 macro Message-ID: <20121126202231.2309E1E0126@mini-me.trailing-edge.com> I suspect the change that broke later versions of macro was related to self redefining macros. The code in question used a complicated scheme to build a hashed symbol table.