On Thursday, December 25, 2014 at 8:52 PM, Jason Stevens wrote:
> I had installed NetBSD 1.2 a while back.. I documented the steps here..
>
> http://gunkies.org/wiki/Installing_NetBSD_1.2_on_the_SIMH_MicroVAX_II
> >
> When NetBSD re-introduced support for the VAX I belive it was mostly the
> 730/750 and then the MicroVAX II. The first bianry distribution I could find
> was 1.2 although it was incredibly unstable in SIMH and I had to use some
> parts of 1.3 to get it installed.
Hmmm... That example is not installing on a MicroVAX II, but on the simh VAX
simulator which simulates the MicroVAX 3900.
Observe:
The ini file contains:
set cpu 32M
The MicroVAX II only supported 16MB of RAM.
The boot log says:
VAX simulator V3.8-1
Loading boot code from ka655x.bin
←[c
KA655-B V5.3, VMB 2.7
Performing normal system tests.
40..39..38..37..36..35..34..33..32..31..30..29..28..27..26..25..
24..23..22..21..20..19..18..17..16..15..14..13..12..11..10..09..
08..07..06..05..04..03..
Tests completed.
A MicroVAX II boot should look something like:
MicroVAX II (KA630) simulator V4.0-0 Beta git commit id:
1d530cc4
sim> b
Loading boot code from ka630.bin
←[c
KA630-A.V1.3
Performing normal system tests.
5..4..3..
Tests completed.
>>>
It isn't surprising that some instabilities under the VAX simulator were noted
since there may have been some new functionality in the 3900 which the NetBSD
1.2 code didn't yet include...
It would be interesting to explore how that version behaved with the MicroVAX
II simulator.
Apart from running under the MicroVAX II or MicroVAX 3900 simulators, Devin
specifically asked about running on the VAX780 simulator.
As it turns out, there were issues with booting and basic installing of NetBSD
on the older VAX systems. Much of NetBSD evolved after the relatively low cost
MicroVAX systems were very common. It seems that although some folks with
older VAX hardware (VAX750, VAX780, VAX730 and VAX8600) may have been running
NetBSD, they didn't install it from CD media, but merely upgraded existing
older BSD operating system installs.
Sometime after Matt Burke implemented the simh simulators for these older VAX
systems and they were included in the simh codebase at github, a few folks
tried to install NetBSD on one of these older systems. It seemed that although
all of these simulators had no problem booting any 'supported' version of VMS
and Ultrix, booting from the NetBSD install CD's didn’t work. On older
systems, Ultrix (and old versions of various BSD operating systems) boot with a
'Boot Block' boot. A 'Boot Block' boot reads sector 0 of the disk into memory
and starts execution at offset 2 of the buffer which had been read. For the
older systems, a Boot Block boot was requested by setting bit 3 of R5 when VMB
was started (i.e. BOOT dev /R5:8). Boot block booting works just find for
Ultrix on these older platforms, but NetBSD's CD distribution media no longer
booted, even though those same CD images would boot just fine on the MicroVAX
3900 simulator (as well as the MicroVAX II).
Some digging showed that the NetBSD disks had a sector 0 which was almost
completely 0s (i.e. it didn't have a secondary bootstrap program in sector 0),
while the sector 0 of Ultrix disks certainly had a secondary bootstrap program
in sector 0. The question was: "Why did these CD images boot on a MicroVAX II
and MicroVAX 3900?". Well, the difference here is the fact that the all of the
MicroVAX systems boot with a version of VMB which is part of the console ROM.
It seems that VMB was extended as it was migrated to the ROMs and several
enhancements were made:
1) Boot Block Boots were attempted automatically if the boot device did not
contain an ODS2 disk structure.
2) Boot Block Boot logic was merged with the PROM boot support that the ROMs
also supported. The PROM boot support allowed booting from a Qbus ROM which
had a particular signature at the beginning of the ROM. When VMB attempted a
Boot Block boot, it read in sector 0 and if the contents of the data written
had the ROM boot signature, then the same logic which initiated booting from
ROM was performed. The ROM signature contains the 'address' of a secondary
bootstrap program. The address is specified as Offset, Size and Starting
address. Since VMB already has a driver which can read the disk it merely
reads the data at the offset and size indicated and then starts execution at
the specified starting address. The ROM based VMB logic will first try this
ROM based boot, but if the signature doesn't match, then it will merely start
execution at offset 2 of the sector 0 data, and thus automatically perform the
legacy Boot Block behavior.
It seems that since the only active development and installation testing for
NetBSD was being done on systems with ROM based VMB someone changed the boot
sector setup process to only put the ROM signature in sector 0 and no one ever
tested on the old hardware.
After this issue was discovered, I created a patched version of VMB.EXE for
simh which also implemented the above logic that already existed in the ROM
based VMB systems. The latest code at github includes VMB.EXE which will run
as needed.
Once this issue was understood by the NetBSD folks someone went back and fixed
the pieces which populate sector 0 on boot disks to both contain the ROM based
signature information AND the legacy boot sector code which had gone missing...
Someone may have also gone back and changed the boot sector data on the
available CD images. This would have allowed booting of those images with any
version of VMB.EXE (possibly with /R5=8), however the program which populates
the boot sector of a newly formatted disk for all of these older installations
also would need to change in order to boot from a disk which was installed from
one of those CD images.
Johnny Billquist was the person who first encountered this problem while
working with the VAX8600 simulator. He was instrumental in engaging and
interacting with the right NetBSD folks to understand what had been broken and
to ultimately fix the current NetBSD VAX code.
As an answer to Devin's original question, I don't think that the generic
NetBSD recipes which exist for installing on the MicroVAX 3900 simulator are
substantially different when using a VAX780 simulator beyond using the latest
simh code from github.
Devin mentioned: http://www.netbsd.org/ports/vax/emulator-howto.html
The instructions there have a configuration file for the MicroVAX 3900
simulator which contains:
load -r /usr/pkg/share/simh/ka655x.bin
set cpu 64m
set rq0 ra92
at rq0 netbsd.dsk
set rq1 cdrom
at rq1 /path/to/vaxcd.iso
at xq0 name-of-network-interface-on-host
boot cpu
For a VAX 780 the following should be equivalently useful:
set cpu 64m
set rq0 ra92
at rq0 netbsd.dsk
set rq1 cdrom
at rq1 /path/to/vaxcd.iso
set xu0 enable
at xu0 name-of-network-interface-on-host
boot rq1
- Mark
_______________________________________________
Simh mailing list
[email protected]
http://mailman.trailing-edge.com/mailman/listinfo/simh
