[-- Attachment #1: Type: text/plain, Size: 3253 bytes --] For anyone that is interested, there is 2 files on Kirk’s DVD that don’t appear on the CD’s mach.86-accent mach.86 The smaller mach.86-accent is a few months newer than the other, and is strictly the kernel. mach.86 contains stuff like the libraries for mach, bindings for pascal, along with an updated libc, and various binaries to run under 4.3BSD. It appears that the Mach project at that time was pretty much in step with the CSRG release. Speaking of pascal, the early version of MIG is actually written in pascal. There is quite a #ifdef ACCENT stuff in the code As well. So the bindings are more than something superficial. I had a major issue trying to use RA81 disks on SIMH, although switching to RP06’s seemed to have made things a little more stable, the larger issue seems to have been the async I/O code, and disabling that increased stability and reduced disk corruption greatly. Setting up the build involved copying files from the ‘cs’ directory to their respective homes, along with the ‘mach/bin/m*’ commands to the /bin directory. Configuring the kernel is very much like a standard BSD kernel config, however the directory needs to exist beforehand, and instead of the in path config command run the config command in the local directory. I have been able to self host a kernel, and build a good portion of world before I realized that the I/O was probably what I was Fighting and went back and restored the 4.3 tape back onto the HP’s and just re-built the kernel to verify it works. For those Wanting the command for SIMH it’s simply ‘set noasync’. The XU adapter worked out of the box with a simple: set xu ena att xu nat:tcp=42323:10.0.2.15:23 Which allowed me to telnet into the VAX, making things much easier than dealing with the console. While this kernel does have mentions of multi processor support I haven’t quite figured out what models (if any) are supported On the VAX, and if SIMH emulates them. While http://www.oboguev.net/vax_mp/ has a very interesting looking multiprocessor VAX Emulation it’s a fictional model based on the microvax, which I’m pretty sure 4.3BSD/Mach’86 is far too old for. And for those who like the gratuitious dmesg, this is a self hosted Mach build loading hp(0,0)boot Boot : hp(0,0)vmunix 393480+61408+138472 start 0x1fa5 Vax boot: memory from 0x92000 to 0x800000 Kernel virtual space from 0x80000000 to 0x82000000. Mach/4.3/2/1 #1: compiled in /usr/mk/MACH on wb2.cs.cmu.edu at Mon Oct 20 12:54:42 1986 physical memory = 8.00 megabytes. available memory = 5.86 megabytes. using 408 buffers containing 0.79 megabytes of memory VAX 11/780, serial#1234(0), hardware ECO level=7(0) mcr0 at tr1 mcr1 at tr2 uba0 at tr3 zs0 at uba0 csr 172520 vec 224, ipl 15 ts0 at zs0 slave 0 dz0 at uba0 csr 160100 vec 300, ipl 15 de0 at uba0 csr 174510 vec 120, ipl 15 de0: hardware address 08:00:2b:0d:d1:48 mba0 at tr8 hp0 at mba0 drive 0 hp1 at mba0 drive 1 hp2 at mba0 drive 2 hp3 at mba0 drive 7 Changing root device to hp0a I uploaded my SIMH config, along with the RP06 disk images here: https://sourceforge.net/projects/bsd42/files/4BSD%20under%20Windows/v0.4/Mach86.zip/download [-- Attachment #2: Type: text/html, Size: 6064 bytes --]
[-- Attachment #1: Type: text/plain, Size: 864 bytes --] On Tue, Jul 16, 2019, at 02:33, Jason Stevens wrote: > I uploaded my SIMH config, along with the RP06 disk images here: https://sourceforge.net/projects/bsd42/files/4BSD%20under%20Windows/v0.4/Mach86.zip/download `ps` is a little unhappy. Hmm. Wonder why. myname# ps aux USER PID %CPU %MEM SZ RSS TT STAT TIME COMMAND ps: cant read u for pid -26328 from /dev/drum ps: cant read indir pte to get u for pid 0 from /dev/kmem ps: cant read u for pid -15008 from /dev/drum ps: cant read indir pte to get u for pid 0 from /dev/kmem ps: cant read u for pid -32767 from /dev/drum ps: cant read indir pte to get u for pid 0 from /dev/kmem ps: cant read u for pid 0 from /dev/drum ps: cant read u for pid 0 from /dev/drum ps: cant read indir pte to get u for pid 124 from /dev/kmem I don't think those PIDs should be negative. :) -- Madeline Autumn-Rose b4@gewt.net [-- Attachment #2: Type: text/html, Size: 1998 bytes --]
[-- Attachment #1: Type: text/plain, Size: 596 bytes --] On Tue, Jul 16, 2019, at 02:33, Jason Stevens wrote: > Emulation it’s a fictional model based on the microvax, which I’m pretty sure 4.3BSD/Mach’86 is far too old for. There are strings in the kernel for the MicroVAX I/II: VAX 8600, serial#%d(%d), hardware ECO level=%d(%d) VAX 820X, urev=%d, patch rev=%d, cpu rev=%d VAX 11/78%c, serial#%d(%d), hardware ECO level=%d(%d) VAX 11/750, hardware level=%d, microcode level=%d MicroVAX-I, hardware level=%d, %c-Float microcode level=%d MicroVAX-II (with%s FPU, SYSCODE:%d, ROM Rev:%d) -- Madeline Autumn-Rose b4@gewt.net [-- Attachment #2: Type: text/html, Size: 1558 bytes --]
[-- Attachment #1: Type: text/plain, Size: 1791 bytes --] Mach has its own ps. The problem is that it just dumps core. I rebuilt it from inside the source, and it has some weird pathing issues, but it dumps core too. At that point I kind of gave up. One fun thing is searching the boot string reveals exactly one machine running this build. Although I would imagine that they probably have nothing left of this.. #N thrash #S DEC VAX 11/785; Mach/4.3/2/1 #O Princeton University, Department of Computer Science #C Pat Parseghian #E princeton!pep #T +1 609 452 6261 #P Engineering Quadrangle, Princeton, NJ 08544 #L 40 21 N / 74 40 W #W princeton!pep (Pat Parseghian); Mon Nov 6 18:19:19 EST 1989 Get Outlook for Android On Wed, Jul 17, 2019 at 3:37 AM +0800, "Madeline Autumn-Rose" <b4@gewt.net> wrote: On Tue, Jul 16, 2019, at 02:33, Jason Stevens wrote: I uploaded my SIMH config, along with the RP06 disk images here: https://sourceforge.net/projects/bsd42/files/4BSD%20under%20Windows/v0.4/Mach86.zip/download `ps` is a little unhappy. Hmm. Wonder why. myname# ps aux USER PID %CPU %MEM SZ RSS TT STAT TIME COMMAND ps: cant read u for pid -26328 from /dev/drum ps: cant read indir pte to get u for pid 0 from /dev/kmem ps: cant read u for pid -15008 from /dev/drum ps: cant read indir pte to get u for pid 0 from /dev/kmem ps: cant read u for pid -32767 from /dev/drum ps: cant read indir pte to get u for pid 0 from /dev/kmem ps: cant read u for pid 0 from /dev/drum ps: cant read u for pid 0 from /dev/drum ps: cant read indir pte to get u for pid 124 from /dev/kmem I don't think those PIDs should be negative. :) -- Madeline Autumn-Rose b4@gewt.net [-- Attachment #2: Type: text/html, Size: 5008 bytes --]
On 2019-07-16 11:32, Jason Stevens wrote:
> While this kernel does have mentions of multi processor support I
> haven’t quite figured out what models (if any) are supported
>
> On the VAX, and if SIMH emulates them. While
> http://www.oboguev.net/vax_mp/ has a very interesting looking
> multiprocessor VAX
The only one I can think of is the 11/782 in this time frame.
The next experiment, the 3520/3540 were much later, and probably slower ;-)
Cheers
[-- Attachment #1: Type: text/plain, Size: 810 bytes --] On Wed, Jul 17, 2019 at 3:56 AM emanuel stiebler <emu@e-bbes.com> wrote: > On 2019-07-16 11:32, Jason Stevens wrote: > > > While this kernel does have mentions of multi processor support I > > haven’t quite figured out what models (if any) are supported > > Historically, the Mach MP work was done on an Encore system not Vaxen but I've forgotten which model and what the processor was. I wanted to say Encore used the NS 16032, but I'm not willing to go out on a limb on that thought. Bob Baron, Mike Accetta or Rich Rashid probably remembers. You might look at the code for Encore support, and that might give you a hint. FWIW: I've lost track of Bob who was at Pitt last time I knew, but Mike is still at Microsoft. I believe I heard Rich retired, but I'm not sure of that. Clem [-- Attachment #2: Type: text/html, Size: 1419 bytes --]