Hi Jordi,
Have you tried running the DEC-10 diagnostics? I'm not a -10 guy's as such, so I
might be wrong here, but I think they are called the KLAD pack.
Search for KLAD on
http://pdp-10.trailing-edge.com/. It should be much easier to pin point
which part of the microcode is causing the problems.
Regards, Tim.
Sent from my Sony Ericsson Xperia arc
Jordi Guillaumes i Pons <jg at jordi.guillaumes.name> wrote:
El 28/11/2012, a les 22:48, Jordi Guillaumes i Pons <jg at jordi.guillaumes.name> va
escriure:
Al 28/11/12 22:47, En/na Paul_Koning at
Dell.com ha escrit:
Is Pi big endian? Does KLH10 assume little endian?
Pi is little endian. And, to add some weirdness to this, I have just completed a TOPS-20
boot using the very same binaries...
What I have found so far:
- The ARM compile of KLH10 does not handle correctly the load of images in DEC-PEXE
format. If the simulator is compiled without optimizations (i.e., -O0 in Makefile) then it
does not spit any error message when it tries to load KLBOOT.EXE but it doesn't work
either. For some reason it does not set up the initial address (but probably the problem
is beyond that). I have been able to dump from a KLH10 running in a amd64 linux to a *.sav
file and it seems to work... but then it does not recognize the TOPS-10 structures. I will
work on both the LOAD problem and the disk problem a little bit.
- TOPS-20 (PANDA) does boot, but the networking code does some weird things. Basically,
lots of "***BUGCHK IPIBLP*** IPNI input buffer list problem" messages whenever I
try to do something with IP networking. I have googled for that bugcheck, and it is
related to the system being able to get IP buffers when it needs them. It could be a race
condition, or some weird thing related to default word sizes.
I'll keep trying... :)
Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES