Re: [Simh] Booting the vax750 simulator.
The suggested “nm –n” is completely useful. From: Clem Cole [mailto:cl...@ccc.com] Sent: Tuesday, July 7, 2015 11:28 AM To: Henry Bent Cc: Mark Pizzolato - Info Comm; simh@trailing-edge.com Subject: Re: [Simh] Booting the vax750 simulator. Boy the bits in my memory are stale and could be confused, but my memory is that there was a way to wash it though the VMS linker to get such a map. The author the VMS linker sits a few feet from me, I just sent him email asked him if he remembers how we did that. I suspect he will make an appearance in my cube after he reads it asking why I want to know :-) Clem On Mon, Jul 6, 2015 at 5:42 PM, Henry Bent hb...@oberlin.edumailto:hb...@oberlin.edu wrote: I took a quick look at this under Ultrix 4.3, where the same problem occurs. There are Ultrix 4.3 sources, so that makes things somewhat easier. The panic occurs in ubarelse(), which releases uba resources. Unfortunately, as far as I know, the Ultrix linker does not have a way to produce a link map. You can at least see what the function addresses are by using nm -n on the kernel you're running. -Henry On 6 July 2015 at 15:41, Mark Pizzolato - Info Comm m...@infocomm.commailto:m...@infocomm.com wrote: Hi Mattis, I’m working with the disk image you provided. I see the failure on the VAX750 simulator. What is the root password on this disk? Do you know if any sort of link map is available from this kernel build? Have you tried earlier versions of Ultrix on the 750 simulator? As I recall, I think I once came across Ultrix sources for some prior versions. It would seem that the VAX750 would be supported pretty much back at the beginning so, if we can reproduce this failure on a version with sources we’d have more to go on. I initially look at the crash message: sim b rq0 Loading boot code from vmb.exe Ultrixboot - V4.0 Sat Mar 31 04:11:56 EST 1990 Loading (a)vmunix ... Sizes: text = 593304 data = 100864 bss = 320516 Starting at 0x2d4d ULTRIX V4.0 (Rev. 161) System #1: Thu May 20 23:26:51 GMT+0100 1976 real mem = 15728640 avail mem = 12378112 using 384 buffers containing 1572864 bytes of memory VAX 11/750, hardware level = 0x9c, microcode level = 99 mcr0 (MS750) at address 0xf2 mba0 at address 0xf28000 0 mba's not configured mba0 at address 0xf2a000 0 mba's not configured uba0 at address 0xf3 uda0 at uba0 uq0 at uda0 csr 172150 vec 774, ipl 15 dz0 at uba0 csr 160100 vec 300, ipl 15 machine check 2: non-existant reference error sumpar = 2 va = 80001a90 errpc = 0 mdr = 0 smr = 0 rdtimo = 0 tbgpar = 0 cacherr = 0 buserr = 140008 mcesr = 0 pc = 80087c61 psl = 4150008 mcsr= 14 I note that the PC mentioned is: 80087c61. If I do the following: sim BRE 80087c61 sim SET CPU HISTORY=300 sim B RQ0 Execution flows up to the indicated address and if I single step from there, the exception is taken. The instruction which causes the exception is: BISL2 #6001,(R0) And R0 contains: But the instructions executed prior to the failing instruction seems to be walking through some memory structures, which no real clue yet exists as to what it might be trying. Hence, my request for a link map and also to try and reproduce this with an OS version that has source… -Mark From: Simh [mailto:simh-boun...@trailing-edge.commailto:simh-boun...@trailing-edge.com] On Behalf Of Mattis Lind Sent: Sunday, June 21, 2015 1:42 PM To: simh@trailing-edge.commailto:simh@trailing-edge.com Subject: Re: [Simh] Booting the vax750 simulator. 2015-06-21 18:22 GMT+02:00 Mark Pizzolato - Info Comm m...@infocomm.commailto:m...@infocomm.com: Hi there Mattis, Well, what you’re asking for (running model specific diagnostics) has been out of scope for all of the simh VAX simulators. The scope has been to be able to run operating system and application software that ran on these systems. If you had source code for the diagnostic you’re trying to run maybe some understanding of why it fails on both the real hardware and what might be needed in the simulator to support it would be possible…. Unfortunately I don't have the source for the diagnostic. It is a pain and that was one of the reasons for trying the simulator in the first place. But I certainly understand that it is quite different to simulate it to pass a diagnostic which tries to do all sorts of strange things than running a real operating system. As for Ultrix 4.0 on the VAX750 simulator, well I haven’t explored the documentation, but it might be possible that by the time Ultrix 4.0 came along, it never got tested on the older hardware. Does this disk image boot using the VAX780 and VAX8600 simulators? This kernel I built was only microvax II and VAX-11/750 so it won't directly boot on a 11/780 nor a VAX8600. But if I add a line CPU VAX780 and
Re: [Simh] Booting the vax750 simulator.
Interesting, I'd never played with lk for Ultrix. It appears that it comes with VAX C, which is part of the base Ultrix packages. VAX Fortran was definitely a separate, paid addon. lk can indeed produce load maps, with lk -M. It also seems that at some point someone thought this would be a good idea to add to regular ld, but never really got very far - ld -M produces what the manual calls a primitive load map, which is really nothing more than a list of object files being linked. -Henry On 7 July 2015 at 14:57, Clem Cole cl...@ccc.com wrote: Mark, No problem -- Indeed, nm -n has been my friend for years. That said, I was replying to Henry's comment about the lack of a link map/output file for Ultrix. Indeed, my experience over the years is that there are times when a link map can be handy and I remember that Ultrix (as opposed to pure joy could create them). I faintly remember when the feature request came into Unix-land asking that the Unix ld be changed to add them and the reply was to use the VMS linker that had been modified to grok a.out (for Fortran - I had forgotten why it was done - Paul reminded me - at the time it was felt that it was easier to port the DEC linker than try to add the features Fortran needed to BSD's ld). I just had forgotten the details, but as I expected; he who authored the hack, did indeed remember them. The key is that if you have lk from the Fortran kit, it can be used to create a.out files. That said, I wonder if their are features supported by ld that lk did not (and might be used by the kernel linkage). I suspect lk ld but not a complete super-set. My experience with the language team is that they would have only done enough to make the DEC languages (and in particular) VAX Fortran work. I have faint memories of playing with lk (but had forgotten the command name and anything much else other than it existed). BTW: when he came over he tells me lk was further hacked to support Elf in later version, he started saying well if you used the itanium VMS port you can get a VMS to create linux binaries. - I said that was more than I though people cared about I think for this exercise. Clem On Tue, Jul 7, 2015 at 2:32 PM, Mark Pizzolato - Info Comm m...@infocomm.com wrote: The suggested “nm –n” is completely useful. *From:* Clem Cole [mailto:cl...@ccc.com] *Sent:* Tuesday, July 7, 2015 11:28 AM *To:* Henry Bent *Cc:* Mark Pizzolato - Info Comm; simh@trailing-edge.com *Subject:* Re: [Simh] Booting the vax750 simulator. Boy the bits in my memory are stale and could be confused, but my memory is that there was a way to wash it though the VMS linker to get such a map. The author the VMS linker sits a few feet from me, I just sent him email asked him if he remembers how we did that. I suspect he will make an appearance in my cube after he reads it asking why I want to know :-) Clem On Mon, Jul 6, 2015 at 5:42 PM, Henry Bent hb...@oberlin.edu wrote: I took a quick look at this under Ultrix 4.3, where the same problem occurs. There are Ultrix 4.3 sources, so that makes things somewhat easier. The panic occurs in ubarelse(), which releases uba resources. Unfortunately, as far as I know, the Ultrix linker does not have a way to produce a link map. You can at least see what the function addresses are by using nm -n on the kernel you're running. -Henry On 6 July 2015 at 15:41, Mark Pizzolato - Info Comm m...@infocomm.com wrote: Hi Mattis, I’m working with the disk image you provided. I see the failure on the VAX750 simulator. What is the root password on this disk? Do you know if any sort of link map is available from this kernel build? Have you tried earlier versions of Ultrix on the 750 simulator? As I recall, I think I once came across Ultrix sources for some prior versions. It would seem that the VAX750 would be supported pretty much back at the beginning so, if we can reproduce this failure on a version with sources we’d have more to go on. I initially look at the crash message: sim b rq0 Loading boot code from vmb.exe Ultrixboot - V4.0 Sat Mar 31 04:11:56 EST 1990 Loading (a)vmunix ... Sizes: text = 593304 data = 100864 bss = 320516 Starting at 0x2d4d ULTRIX V4.0 (Rev. 161) System #1: Thu May 20 23:26:51 GMT+0100 1976 real mem = 15728640 avail mem = 12378112 using 384 buffers containing 1572864 bytes of memory VAX 11/750, hardware level = 0x9c, microcode level = 99 mcr0 (MS750) at address 0xf2 mba0 at address 0xf28000 0 mba's not configured mba0 at address 0xf2a000 0 mba's not configured uba0 at address 0xf3 uda0 at uba0 uq0 at uda0 csr 172150 vec 774, ipl 15 dz0 at uba0 csr 160100 vec 300, ipl 15 machine check 2: non-existant reference error sumpar = 2 va = 80001a90 errpc = 0 mdr = 0 smr = 0 rdtimo = 0
[Simh] Booting the vax750 simulator
[...] This kernel I built was only microvax II and VAX-11/750 so it won't directly boot on a 11/780 nor a VAX8600. But if I add a line CPU VAX780 and CPU VAX8600 it boots on both. But still no-go on vax750. I even tried to remove some more optional features in the kernel config but no difference. Still boot on vax780, microvax2 and vax8600 simulator though. Here is the image I tried : https://dl.dropboxusercontent.com/u/96935524/Datormusuem/VAX11-750/ultrix3.dsk [...] I'm getting a dead link here. Bill ___ Simh mailing list Simh@trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh
Re: [Simh] Booting the vax750 simulator.
Boy the bits in my memory are stale and could be confused, but my memory is that there was a way to wash it though the VMS linker to get such a map. The author the VMS linker sits a few feet from me, I just sent him email asked him if he remembers how we did that. I suspect he will make an appearance in my cube after he reads it asking why I want to know :-) Clem On Mon, Jul 6, 2015 at 5:42 PM, Henry Bent hb...@oberlin.edu wrote: I took a quick look at this under Ultrix 4.3, where the same problem occurs. There are Ultrix 4.3 sources, so that makes things somewhat easier. The panic occurs in ubarelse(), which releases uba resources. Unfortunately, as far as I know, the Ultrix linker does not have a way to produce a link map. You can at least see what the function addresses are by using nm -n on the kernel you're running. -Henry On 6 July 2015 at 15:41, Mark Pizzolato - Info Comm m...@infocomm.com wrote: Hi Mattis, I’m working with the disk image you provided. I see the failure on the VAX750 simulator. What is the root password on this disk? Do you know if any sort of link map is available from this kernel build? Have you tried earlier versions of Ultrix on the 750 simulator? As I recall, I think I once came across Ultrix sources for some prior versions. It would seem that the VAX750 would be supported pretty much back at the beginning so, if we can reproduce this failure on a version with sources we’d have more to go on. I initially look at the crash message: sim b rq0 Loading boot code from vmb.exe Ultrixboot - V4.0 Sat Mar 31 04:11:56 EST 1990 Loading (a)vmunix ... Sizes: text = 593304 data = 100864 bss = 320516 Starting at 0x2d4d ULTRIX V4.0 (Rev. 161) System #1: Thu May 20 23:26:51 GMT+0100 1976 real mem = 15728640 avail mem = 12378112 using 384 buffers containing 1572864 bytes of memory VAX 11/750, hardware level = 0x9c, microcode level = 99 mcr0 (MS750) at address 0xf2 mba0 at address 0xf28000 0 mba's not configured mba0 at address 0xf2a000 0 mba's not configured uba0 at address 0xf3 uda0 at uba0 uq0 at uda0 csr 172150 vec 774, ipl 15 dz0 at uba0 csr 160100 vec 300, ipl 15 machine check 2: non-existant reference error sumpar = 2 va = 80001a90 errpc = 0 mdr = 0 smr = 0 rdtimo = 0 tbgpar = 0 cacherr = 0 buserr = 140008 mcesr = 0 pc = 80087c61 psl = 4150008 mcsr= 14 I note that the PC mentioned is: 80087c61. If I do the following: sim BRE 80087c61 sim SET CPU HISTORY=300 sim B RQ0 Execution flows up to the indicated address and if I single step from there, the exception is taken. The instruction which causes the exception is: BISL2 #6001,(R0) And R0 contains: But the instructions executed prior to the failing instruction seems to be walking through some memory structures, which no real clue yet exists as to what it might be trying. Hence, my request for a link map and also to try and reproduce this with an OS version that has source… -Mark *From:* Simh [mailto:simh-boun...@trailing-edge.com] *On Behalf Of *Mattis Lind *Sent:* Sunday, June 21, 2015 1:42 PM *To:* simh@trailing-edge.com *Subject:* Re: [Simh] Booting the vax750 simulator. 2015-06-21 18:22 GMT+02:00 Mark Pizzolato - Info Comm m...@infocomm.com : Hi there Mattis, Well, what you’re asking for (running model specific diagnostics) has been out of scope for all of the simh VAX simulators. The scope has been to be able to run operating system and application software that ran on these systems. If you had source code for the diagnostic you’re trying to run maybe some understanding of why it fails on both the real hardware and what might be needed in the simulator to support it would be possible…. Unfortunately I don't have the source for the diagnostic. It is a pain and that was one of the reasons for trying the simulator in the first place. But I certainly understand that it is quite different to simulate it to pass a diagnostic which tries to do all sorts of strange things than running a real operating system. As for Ultrix 4.0 on the VAX750 simulator, well I haven’t explored the documentation, but it might be possible that by the time Ultrix 4.0 came along, it never got tested on the older hardware. Does this disk image boot using the VAX780 and VAX8600 simulators? This kernel I built was only microvax II and VAX-11/750 so it won't directly boot on a 11/780 nor a VAX8600. But if I add a line CPU VAX780 and CPUVAX8600 it boots on both. But still no-go on vax750. I even tried to remove some more optional features in the kernel config but no difference. Still boot on vax780, microvax2 and vax8600 simulator though.
Re: [Simh] Booting the vax750 simulator.
As I was typing, indeed he came over. The answer is the lk command which is part of the VAX Fortran kit and is the VMS linker. On Tue, Jul 7, 2015 at 2:28 PM, Clem Cole cl...@ccc.com wrote: Boy the bits in my memory are stale and could be confused, but my memory is that there was a way to wash it though the VMS linker to get such a map. The author the VMS linker sits a few feet from me, I just sent him email asked him if he remembers how we did that. I suspect he will make an appearance in my cube after he reads it asking why I want to know :-) Clem On Mon, Jul 6, 2015 at 5:42 PM, Henry Bent hb...@oberlin.edu wrote: I took a quick look at this under Ultrix 4.3, where the same problem occurs. There are Ultrix 4.3 sources, so that makes things somewhat easier. The panic occurs in ubarelse(), which releases uba resources. Unfortunately, as far as I know, the Ultrix linker does not have a way to produce a link map. You can at least see what the function addresses are by using nm -n on the kernel you're running. -Henry On 6 July 2015 at 15:41, Mark Pizzolato - Info Comm m...@infocomm.com wrote: Hi Mattis, I’m working with the disk image you provided. I see the failure on the VAX750 simulator. What is the root password on this disk? Do you know if any sort of link map is available from this kernel build? Have you tried earlier versions of Ultrix on the 750 simulator? As I recall, I think I once came across Ultrix sources for some prior versions. It would seem that the VAX750 would be supported pretty much back at the beginning so, if we can reproduce this failure on a version with sources we’d have more to go on. I initially look at the crash message: sim b rq0 Loading boot code from vmb.exe Ultrixboot - V4.0 Sat Mar 31 04:11:56 EST 1990 Loading (a)vmunix ... Sizes: text = 593304 data = 100864 bss = 320516 Starting at 0x2d4d ULTRIX V4.0 (Rev. 161) System #1: Thu May 20 23:26:51 GMT+0100 1976 real mem = 15728640 avail mem = 12378112 using 384 buffers containing 1572864 bytes of memory VAX 11/750, hardware level = 0x9c, microcode level = 99 mcr0 (MS750) at address 0xf2 mba0 at address 0xf28000 0 mba's not configured mba0 at address 0xf2a000 0 mba's not configured uba0 at address 0xf3 uda0 at uba0 uq0 at uda0 csr 172150 vec 774, ipl 15 dz0 at uba0 csr 160100 vec 300, ipl 15 machine check 2: non-existant reference error sumpar = 2 va = 80001a90 errpc = 0 mdr = 0 smr = 0 rdtimo = 0 tbgpar = 0 cacherr = 0 buserr = 140008 mcesr = 0 pc = 80087c61 psl = 4150008 mcsr= 14 I note that the PC mentioned is: 80087c61. If I do the following: sim BRE 80087c61 sim SET CPU HISTORY=300 sim B RQ0 Execution flows up to the indicated address and if I single step from there, the exception is taken. The instruction which causes the exception is: BISL2 #6001,(R0) And R0 contains: But the instructions executed prior to the failing instruction seems to be walking through some memory structures, which no real clue yet exists as to what it might be trying. Hence, my request for a link map and also to try and reproduce this with an OS version that has source… -Mark *From:* Simh [mailto:simh-boun...@trailing-edge.com] *On Behalf Of *Mattis Lind *Sent:* Sunday, June 21, 2015 1:42 PM *To:* simh@trailing-edge.com *Subject:* Re: [Simh] Booting the vax750 simulator. 2015-06-21 18:22 GMT+02:00 Mark Pizzolato - Info Comm m...@infocomm.com : Hi there Mattis, Well, what you’re asking for (running model specific diagnostics) has been out of scope for all of the simh VAX simulators. The scope has been to be able to run operating system and application software that ran on these systems. If you had source code for the diagnostic you’re trying to run maybe some understanding of why it fails on both the real hardware and what might be needed in the simulator to support it would be possible…. Unfortunately I don't have the source for the diagnostic. It is a pain and that was one of the reasons for trying the simulator in the first place. But I certainly understand that it is quite different to simulate it to pass a diagnostic which tries to do all sorts of strange things than running a real operating system. As for Ultrix 4.0 on the VAX750 simulator, well I haven’t explored the documentation, but it might be possible that by the time Ultrix 4.0 came along, it never got tested on the older hardware. Does this disk image boot using the VAX780 and VAX8600 simulators? This kernel I built was only microvax II and VAX-11/750 so it won't directly boot on a 11/780 nor a VAX8600. But if I add a line CPU VAX780 and CPUVAX8600 it
[Simh] simh-V3.8-1
The only old copy of simh I was able to get was the 3.8-1 version and it involves DECnet. Does anyone know anything about a DECnet for simh or on the internet anywhere? Why is this simh called that? Is it a specialized version by another party other than the maintainer? Bill ___ Simh mailing list Simh@trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh