TOPS-10 boot failure on Raspberry Pi 2

[JohnForecast]

I've built kn10-kl on the latest Stretch version of Raspbian (2018-3-13) for a Raspberry Pi 2. It works perfectly running TOPS-20 but when I try to boot TOPS-10 I get the following error loading the boot loader:

KLH10 2.0k-Rhialto (MyKL) built Mar 22 2018 20:45:27 Copyright ? 2002 Kenneth L. Harrenstien -- All Rights Reserved. This program comes "AS IS" with ABSOLUTELY NO WARRANTY.

Compiled for unknown-linux-gnueabihf on armv7l with word model USEINT Emulated config: CPU: KL10-extend SYS: T20 Pager: KL APRID: 3600 Memory: 8192 pages of 512 words (SHARED) Time interval: INTRP Base: OSGET Interval default: 60Hz Internal clock: OSINT Other: MCA25 CIRC JPC DEBUG PCCACHE CTYINT EVHINT Devices: DTE RH20 RPXX(DP) TM03(DP) NI20(DP) [MEM: Allocating 8192 pages shared memory, clearing...done]

KLH10# ; KLH10 configuration for TOPS-10 test system KLH10# KLH10# ; DTE requires ackdly to avoid T10 race condition KLH10# devdef dte0 200 dte master ackdly=5 KLH10# devdef rh0 540 rh20 KLH10# devdef rh1 544 rh20 KLH10# devdef dsk0 rh0.0 rp dppath=/usr/local/bin/dprpxx type=rp06 sn=4747 format=dbd9 path=T10-RP06.0-dbd9 KLH10# devdef dsk1 rh0.1 rp dppath=/usr/local/bin/dprpxx type=rp06 sn=1026 format=dbd9 path=T10-RP06.1-dbd9 KLH10# devdef mta0 rh1.0 tm03 dppath=/usr/local/bin/dptm03 type=TU77 KLH10# KLH10# ; NI: param "c3dly" to avoid T10 race. KLH10# ; NI: param "rdtmo" to avoid system hangups with OSF/1 V3.0 KLH10# ; KLH10# ; Decnet node TWONKY, 9.429 KLH10# devdef ni0 564 ni20 dppath=/usr/local/bin/dpni20 dedic=true ifmeth=tap+bridge ifc=tap1 decnet=1 doarp=0 enaddr=aa:00:04:00:1c:fc dpdelay=12 c3dly=2 rdtmo=3 KLH10# KLH10# load klboot.exe Using word format "c36"... Loading aborted, read failed for file page 1, proc page 0 Loading aborted, read failed for file page 2, proc page 448 Loading aborted, read failed for file page 4, proc page 451 Loaded "klboot.exe": Format: DEC-PEXE Data: 0, Symwds: 0, Low: 01000000, High: 0, Startaddress: 0 Entvec: 0 wds at 0 KLH10# [EOF on klt10.ini]

[larsbrinkhoff]

It looks like KLH10 didn't quite like the format of your klboot.exe file. "Loading aborted" doesn't look good, and for sure Data, Low, High, Startaddress, etc is wrong.

Some guesses:

• The contents of the file is messed up.
• The contents are fine, but the format is wrong.

[larsbrinkhoff]

I'm not a TOPS-10 user myself, so I can't offer much more. If you find someone else using KLH10 to run TOPS-10, they might be able to help.

There's also SIMH that can run the KS10 version of TOPS-10.

[JohnForecast]

I just compared klboot.exe with an older version of the klh10 distribution (2.0h, I think) and there's no difference. Maybe I'll build a debug version and see what's happening.

I've been using SIMH for running TOPS-10 but I had an issue get the DZ-11 configured correctly (now solved) and was hoping to move to KLH10 so I could use LAT terminals.

[Rhialto]

Did you try it on another cpu architecture? Maybe that makes the difference. I have not knowingly modified anything in loading files, just mainly networking stuff.

[JohnForecast]

I built it on MacOS Sierra and there we no problems loading klboot.exe. It does seem to be architecture dependent - I'm building a debug version to see if I can get more information.

[jguillaumes]

El 6 abr 2018, a les 19:44, JohnForecast notifications@github.com va escriure:

I built it on MacOS Sierra and there we no problems loading klboot.exe. It does seem to be architecture dependent - I'm building a debug version to see if I can get more information.

This sounds familiar to me...

Try building the simulator without compiler optimizations (you’ll need to hack the makefile a little bit).

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

[JohnForecast]

I tried building with optimizations, it got rid of the error messages but nothing was loaded into memory. I even tried building with clang but the same problem. It almost seems as if there's some memory corruption happening. If I just try to load "klboot.exe", I get the error messages as before. If I enable tracing of the load process, I get the following output:

KLH10# set ld_debug=1 ld_debug: Off => On KLH10# load klboot.exe Using word format "c36"... DIR section = 012 wds No entvec block, word: 01777,,01 Page 01: file page 0 (n=1.) Page 02: file page 0700 (n=2.) Page 0702: clear Page 04: file page 0703 (n=43.) Page 0756: clear Loaded "klboot.exe": Format: DEC-PEXE Data: 0, Symwds: 0, Low: 01000000, High: 0, Startaddress: 0 Entvec: 0 wds at 0

Nothing was loaded but where did the error messages go?

[JohnForecast]

I have an old 2.0h version of KLH10 that I modified to run on the original Raspberry Pi back in 2014. I just built it on the same system where I was having the boot problems - it is able to load "klboot.exe" successfully. If it's of any use, I can make a tar file available along with notes I made at the time about the changes that were required.

[JohnForecast]

I've tracked down the problem. There was a difference between sizeof(wfoff_t) in wfio.c (8) and feload.c (4) which results in wf_seek() failing silently. The TOPS-20 boot loading code never calls wf_seek() so this problem never showed up.

The core problem is that the system include files typedef off_t as "long" unless _FILE_OFFSET_BITS == 64 in which case it will be defined as "long long". The simplest fix is to make sure that cenv.h is included first in all source files - it works OK in my Raspberry Pi environment but I haven't tested any other environment.

[larsbrinkhoff]

That's great! Would you like to submit a pull request?

[Rhialto]

Good catch! In my view, everything in cenv.h is a candidate for the configure script. I just hadn't gotten around to it any further...

[robmayes69]

Does anyone know if this issue is resolved in recent builds as I'm getting HALT messages when using a Raspberry Pi compile of SimH4.

[larsbrinkhoff]

@robmayes69, I don't think this was fixed.

[JohnForecast]

The change referenced above was to the KLH10 emulator which has nothing to do with SimH4. I just checked and this change is still present in the KLH10 repository.

[Rhialto]

I was about to create a merge request based on https://github.com/PDP-10/klh10/issues/31#issuecomment-379594901 but both files (wfio.c and feload.c) already #include "cenv.h". It's also included from "klh10.h" which is included from lots of other files. The only files that don't include either of those are, if I filtered correctly,

osdnet.c
prmstr.c
tapedd.c
udlconv.c
uexbconv.c
wfconv.c
wxtest.c

[Rhialto]

So I went through all files and made sure that each of them include either cenv.h or klh10.h as the very first include. See MR #62.

In fact I would like to get rid of the whole cenv.h file altogether. The one or two items that may still be relevant (and as far as I can see that is only Long File Support and termios vs. BSD tty) should be put into the configure script. Everything then only needs to #include "config.h" as first.

There is also still #59 which cleans up stuff from the old build system that likely doesn't work any more. If nobody can convince me to keep it, it will go soon.