b-em - bug or misunderstanding

want to talk about MESS/model b/beebem/b-em/electrem/elkulator? do it here!
Post Reply
User avatar
pau1ie
Posts: 709
Joined: Thu May 10, 2012 9:48 pm
Location: Bedford
Contact:

b-em - bug or misunderstanding

Post by pau1ie » Fri Sep 06, 2019 9:27 pm

Playing with b-em (Current master) I find the following.

*ADFS doesn't work. - It looks for a file. *FADFS works OK.

I can't work out how to load a hard disc image, or indeed if you can't load one, where it expects to find one (I am on Arch Linux)

When firing up Panos, I loaded the welcome disc in drive 1 leaving the start up disc in drive 0. It wouldn't load the welcome program, and when I did a cat :0 and :1, the disc catalogue is the same. I had a look at the disc images, and they look correct.

Panos doesn't seem to work on a model B - It needs a master 128, but maybe this is just the version included.

I am happy to raise bugs, but I wanted to check I wasn't missing something obvious.

User avatar
geraldholdsworth
Posts: 529
Joined: Tue Nov 04, 2014 9:42 pm
Location: Inverness, Scotland
Contact:

Re: b-em - bug or misunderstanding

Post by geraldholdsworth » Sat Sep 07, 2019 9:09 am

pau1ie wrote:
Fri Sep 06, 2019 9:27 pm
Playing with b-em (Current master) I find the following.

*ADFS doesn't work. - It looks for a file. *FADFS works OK.
That is also the behaviour of a real machine. *ADFS (or A-Break) will attempt to mount a directory, but *FADFS (or F-Break) will not. On a Master, IIRC, it is configurable whether ADFS mounts a directory or not.
Gerald Holdsworth
Repton Resource Page
www.reptonresourcepage.co.uk

User avatar
jgharston
Posts: 3696
Joined: Thu Sep 24, 2009 11:22 am
Location: Whitby/Sheffield
Contact:

Re: b-em - bug or misunderstanding

Post by jgharston » Sat Sep 07, 2019 9:24 am

pau1ie wrote:
Fri Sep 06, 2019 9:27 pm
Panos doesn't seem to work on a model B - It needs a master 128, but maybe this is just the version included.
I've had PanOS working on an Electron, so it should work on a BBC.

Code: Select all

$ bbcbasic
PDP11 BBC BASIC IV Version 0.25
(C) Copyright J.G.Harston 1989,2005-2015
>_

Coeus
Posts: 1382
Joined: Mon Jul 25, 2016 11:05 am
Contact:

Re: b-em - bug or misunderstanding

Post by Coeus » Sat Sep 07, 2019 2:09 pm

pau1ie wrote:
Fri Sep 06, 2019 9:27 pm
I can't work out how to load a hard disc image, or indeed if you can't load one, where it expects to find one (I am on Arch Linux)
The supplied ROMs all assume SCSI hard discs as per the original winchester so you would need SCSI hard discs enabled from the Disc menu then on Linux It looks in ~/.config/b-em/scsi for four pairs of files named scsi0.dat, scsi0.dsc etc. up to 3. On windows it looks in a B-Em sub-directory of the users' data directory. It does also support IDE discs where the bytes actually used are interleaved with zeros - you would need to load patched ADFS to use that.
pau1ie wrote:
Fri Sep 06, 2019 9:27 pm
When firing up Panos, I loaded the welcome disc in drive 1 leaving the start up disc in drive 0. It wouldn't load the welcome program, and when I did a cat :0 and :1, the disc catalogue is the same. I had a look at the disc images, and they look correct.
I'll check this out. I am sure I was able to load PanOS from floppy to hard disc in the past but it was a while back. Are these ADFS or DFS floppies?

User avatar
CMcDougall
Posts: 6795
Joined: Wed Feb 02, 2005 3:13 pm
Location: Shadow in a Valley of Scotland
Contact:

Re: b-em - bug or misunderstanding

Post by CMcDougall » Sat Sep 07, 2019 7:23 pm

/\ I had it working, so must be DFS as ADFS does not exist in any of my Beebs / elks :D

If I mind rightly, I wrote them all with UPURS, worked a treat from Jon's site (DFS / ADFS explorer author)
ImageImageImage

User avatar
pau1ie
Posts: 709
Joined: Thu May 10, 2012 9:48 pm
Location: Bedford
Contact:

Re: b-em - bug or misunderstanding

Post by pau1ie » Sat Sep 07, 2019 7:24 pm

Coeus wrote:
Sat Sep 07, 2019 2:09 pm
Are these ADFS or DFS floppies?
It's the images that come with the emulator. They are DFS.
Coeus wrote:
Sat Sep 07, 2019 2:09 pm
on Linux It looks in ~/.config/b-em/scsi for four pairs of files
I expect these need to have been sourced from elsewhere.

I was expecting to use IDE as that is what is in my datacentre, but as you say, I am using patched ADFS to access that. Thanks. I will hopefully get round to writing this up.

Coeus
Posts: 1382
Joined: Mon Jul 25, 2016 11:05 am
Contact:

Re: b-em - bug or misunderstanding

Post by Coeus » Sun Sep 08, 2019 7:13 pm

pau1ie wrote:
Fri Sep 06, 2019 9:27 pm
When firing up Panos, I loaded the welcome disc in drive 1 leaving the start up disc in drive 0. It wouldn't load the welcome program, and when I did a cat :0 and :1, the disc catalogue is the same. I had a look at the disc images, and they look correct.
This appears to be a Panos thing. I get the same result as you that 'cat :1' in Panos catalogues the disc in drive 0 but if I turn off the virtual 32016 by delecting it in the Tube menu and then catalogue the two discs from the BASIC prompt they catalogue correctly as two distinct discs.
pau1ie wrote:
Fri Sep 06, 2019 9:27 pm
Panos doesn't seem to work on a model B - It needs a master 128, but maybe this is just the version included.
Is it simply that the BBC B does not have an ADFS ROM?
pau1ie wrote:
Sat Sep 07, 2019 7:24 pm
Coeus wrote:
Sat Sep 07, 2019 2:09 pm
on Linux It looks in ~/.config/b-em/scsi for four pairs of files
I expect these need to have been sourced from elsewhere.
That's one option. Another is to use the hdfmt program included in the b-em distribution. This is a command-line tool to create a blank SCSI hard disc. It takes the base name of the image file and a size, creates the .dat and .dsc files and puts an empty ADFS filesystem of the correct size on the disc image so it is then ready to mount.

User avatar
pau1ie
Posts: 709
Joined: Thu May 10, 2012 9:48 pm
Location: Bedford
Contact:

Re: b-em - bug or misunderstanding

Post by pau1ie » Sun Sep 08, 2019 10:08 pm

Coeus wrote:
Sun Sep 08, 2019 7:13 pm
This appears to be a Panos thing.
I was following the Cambridge Co-procesor user guide linked from http://chrisacorns.computinghistory.org ... i/Sci.html section 3.2.3 which says with the welcome disc in drive 1 I can type

Code: Select all

welcome
and run the welcome programs. I don't know if it is an error in panos or b-em, but it doesn't work how the manual says it should. Also further along in the same section it says I should be able to

Code: Select all

cat :1
to see the contents of that disc drive.
Coeus wrote:
Sun Sep 08, 2019 7:13 pm
Is it simply that the BBC B does not have an ADFS ROM?
I don't think so because Panos uses DFS. It doesn't error, it just hangs.
Coeus wrote:
Sun Sep 08, 2019 7:13 pm
the hdfmt program included in the b-em distribution.
Ah - thanks - I hadn't noticed that. I will have to write this all up on the wiki.

Coeus
Posts: 1382
Joined: Mon Jul 25, 2016 11:05 am
Contact:

Re: b-em - bug or misunderstanding

Post by Coeus » Tue Sep 10, 2019 12:36 am

I did discover a bug in the latest master where the 32016 would not give the Pandora prompt which was due to failing to set the clock speed of the 32016 fast enough. I have fixed that and pushed the fix to master on GitHub.

As for the case where cat :1 does not do as expected, this does seem to be firmly a Panos issue. Despite what the documentation seems to be claiming, when I try this and look at what happens over the tube, the interesting bit is this:

Code: Select all

10/09/2019 01:00:29 DEBUG tube: parasite write R2=08
10/09/2019 01:00:29 DEBUG tube: host read R2=08
10/09/2019 01:00:29 DEBUG tube: parasite write R2=7F
10/09/2019 01:00:29 DEBUG tube: host read R2=7F
10/09/2019 01:00:29 DEBUG tube: parasite write R2=10
10/09/2019 01:00:29 DEBUG tube: host read R2=10
10/09/2019 01:00:29 DEBUG tube: parasite write R2=00
10/09/2019 01:00:29 DEBUG tube: host read R2=00
10/09/2019 01:00:29 DEBUG tube: parasite write R2=00
10/09/2019 01:00:29 DEBUG tube: host read R2=00
10/09/2019 01:00:29 DEBUG tube: parasite write R2=00
10/09/2019 01:00:29 DEBUG tube: host read R2=00
10/09/2019 01:00:29 DEBUG tube: parasite write R2=00
10/09/2019 01:00:29 DEBUG tube: host read R2=00
10/09/2019 01:00:29 DEBUG tube: parasite write R2=00
10/09/2019 01:00:29 DEBUG tube: host read R2=00
10/09/2019 01:00:29 DEBUG tube: parasite write R2=0F
10/09/2019 01:00:29 DEBUG tube: host read R2=0F
10/09/2019 01:00:29 DEBUG tube: parasite write R2=22
10/09/2019 01:00:29 DEBUG tube: host read R2=22
10/09/2019 01:00:29 DEBUG tube: parasite write R2=00
10/09/2019 01:00:29 DEBUG tube: host read R2=00
10/09/2019 01:00:29 DEBUG tube: parasite write R2=00
10/09/2019 01:00:29 DEBUG tube: host read R2=00
10/09/2019 01:00:29 DEBUG tube: parasite write R2=53
10/09/2019 01:00:29 DEBUG tube: host read R2=53
10/09/2019 01:00:29 DEBUG tube: parasite write R2=03
10/09/2019 01:00:29 DEBUG tube: host read R2=03
10/09/2019 01:00:29 DEBUG tube: parasite write R2=00
10/09/2019 01:00:29 DEBUG tube: host read R2=00
10/09/2019 01:00:29 DEBUG tube: parasite write R2=0F
10/09/2019 01:00:29 DEBUG tube: host read R2=0F
10/09/2019 01:00:29 DEBUG tube: parasite write R2=DF
10/09/2019 01:00:29 DEBUG tube: host read R2=DF
10/09/2019 01:00:29 DEBUG tube: parasite write R2=46
10/09/2019 01:00:29 DEBUG tube: host read R2=46
10/09/2019 01:00:29 DEBUG tube: parasite write R2=00
10/09/2019 01:00:29 DEBUG tube: host read R2=00
10/09/2019 01:00:29 DEBUG tube: parasite write R2=10
Decoding this with reference to: http://mdfs.net/Software/Tube/Docs/Protocol the &08 written to R2 starts an OSWORD call aith A=&7F, i.e. read sectors from floppy. The parameter block is backwards so this is reading drive 0, track 0, sector 0, for 2 sectors. So it seems Panos is generating the catalogue listing rather than asking DFS to do it, and it supplying drive 0 even when :1 is specified on the command line.
Last edited by Coeus on Tue Sep 10, 2019 12:38 am, edited 1 time in total.

User avatar
pau1ie
Posts: 709
Joined: Thu May 10, 2012 9:48 pm
Location: Bedford
Contact:

Re: b-em - bug or misunderstanding

Post by pau1ie » Fri Sep 13, 2019 10:38 pm

Coeus wrote:
Tue Sep 10, 2019 12:36 am
I did discover a bug
Thanks for investigating this and fixing it so quickly. I can confirm I can see the panos prompt in both master and model B now. In the hope of getting panos to work on floppies I tried downloading the images for version 1.1 from mdfs.net. That didn't work and is probably related to the error messages I got starting:

Code: Select all

[pc=329] Writing outside of RAM @ ECD664 = 00
The memory address increments to FFFFFF, then becomes slightly more random. I am going to leave this alone because I have decided to limit myself to documenting how to use the disc images that are supplied with b-em.

I had some issues with hdfmt, but raised issue 90 for that.

Thanks for your work on b-em, it was already a great piece of software, and your work is steadily improving it!
Last edited by pau1ie on Fri Sep 13, 2019 10:42 pm, edited 1 time in total.

Coeus
Posts: 1382
Joined: Mon Jul 25, 2016 11:05 am
Contact:

Re: b-em - bug or misunderstanding

Post by Coeus » Sat Sep 14, 2019 9:43 pm

pau1ie wrote:
Fri Sep 13, 2019 10:38 pm

Code: Select all

[pc=329] Writing outside of RAM @ ECD664 = 00
That's interesting as I have just been trying to work out my Master with MOS 3.52 b-em configuration would crash with a similar error message and yet the MOS 3.50 didn't. Likewise with 3.22 and 3.20. These are patched versions for Y2K and various bugs. After bisecting the changes and finding it always crashed, it turned out to be nothing to do with the OS ROM at all. Instead, I had been experimenting with the PDP11 that I recently added (from PiTubeDirect) and had a copy of PDP11 BASIC loaded as ROM on the 3.52 and 3.22 models. I suspect that maybe being transferred to the 32016 as the default language thus crashing the 32016 as it is PDP11 code, not 32016.

You may have read in the Issue 90 bug report that I discovered another issue, this time with the floppy disc format selection, where it was treating 2nd and subsequent DOSPlus/GEM discs as double, but sequential sided whereas the image files are actually interleaved sides (like DSD). That caused the GEM/DOSPlus install to fail.

So I have now been through the DOS/GEM install so I will try the PanOS one see if it works for me.

Coeus
Posts: 1382
Joined: Mon Jul 25, 2016 11:05 am
Contact:

Re: b-em - bug or misunderstanding

Post by Coeus » Wed Sep 18, 2019 8:18 pm

I cannot get Panos to install from floppies either - it always crashes. I have opened bug https://github.com/stardot/b-em/issues/91

In the mean time, if you would like a hard disc image with it pre-installed I could supply one.

Coeus
Posts: 1382
Joined: Mon Jul 25, 2016 11:05 am
Contact:

Re: b-em - bug or misunderstanding

Post by Coeus » Fri Sep 20, 2019 12:16 am

Issue 91 turned out to be a tube flow-control problem causing overrun and corruption on saving to ADFS, but probably only from the 32016. That is now fixed so Pandora/Panos things should work much better. Installing to ADFS SCSI hard disc from the floppies now works with the current master.

User avatar
pau1ie
Posts: 709
Joined: Thu May 10, 2012 9:48 pm
Location: Bedford
Contact:

Re: b-em - bug or misunderstanding

Post by pau1ie » Fri Sep 27, 2019 10:06 pm

Thanks for your work investigating and fixing bugs. I will have another go at this when I get the urge!

Post Reply