New MAME release

[+dhe]

I think I have found a bug!?!

 

I start MAME with:

D:\mame>mame genmod -bios 1.00 -peb:slot3 horizon -peb:slot5 tirs232 -peb:slot8 hfdc -peb:slot8:hfdc:h1 generic -peb:slot8:hfdc:h2 generic -peb:slot8:hfdc:h3 generic -peb:slot8:hfdc:f3 525dd -peb:slot8:hfdc:f4 525dd -hard1 genos7boot.HD -hard2 Bootdisk2.HD -hard3 Bootdisk3.HD -flop1 DSDD1.dsk -flop2 DSDD2.hfe -flop3 DSDD3.dsk -flop4 DSDD4.dsk -serl1 socket.localhost:10000

 

GeneveOS 7.3 boots up.

 

If I do a: dir h: (scsi1) I get device error.

If I do a: dir i: (scsi2) things lock up.

 

What tools are available to see if this is a bug in MAME or a bug in GeneveOS?

 

[+9640News]

What is your SCSMAP settings?

[+mizapf]

The same happens for me. I'd vote for a GeneveOS bug.

 

If you start the debugger (command line option "-debug") you see that it is spinning in these lines when you do a "dir i:" or "dir j:":

 

0d24 movb @>4fe2,R2
0d28 andi R2,>4000
0d2c jeq >0d24

 

Maybe we can locate that in the source code.

 

Edit: This looks like waiting for the SCSI arbitration-in-progress bit. Interestingly, SCS1 does not lead into this loop.

[+mizapf]

2 minutes ago, 9640News said:

What is your SCSMAP settings?

It seems as if he does not use the SCSI card.

 

I have these scsmappings:

 

SCS1 = 0

SCS2 = 1

SCS3 = 2

SCS4 = 8

SCS5 = 9

SCS6 = A

[+InsaneMultitasker]

There might not be a device error trap to detect whether a SCSI card is present before performing the low level IO.  Long ago, a DSR EPROM was not required to use the SCSI card with a Geneve.  I think we've moved beyond that and if validation hasn't already been added to 7.3x as part of the in-process cleanup, the code should be added before a next release.  Can someone please add this issue to the Geneve OS development thread if it is not a MAME issue?

[+mizapf]

MAME 0.243 has been released today. Luckily, my recent works are included already, that is, async mode for TIPI, and the details about the counter behaviour of the 9901.

[+dhe]

Coming to Whtech soon (please)?

[+mizapf]

There you are. Just uploaded.

[+9640News]

@mizapf

 

I am not sure if I am trying something presently unsupported with the TIPI and MAME.  Using the Geneve emulation.  I am able to access directory listings on the PI without issue.

 

I have both the embedded telnet client in MDOS as well as MyTerm for the TIPI.  When I attempt to use the TCP services, MAME aborts without any error messages and closes out the emulation.  

 

I guess my first question, is if TCP services are working, does Matt's telnet client work?

 

Beery

[+mizapf]

I'd need some more detailed descriptions to test it. I have certainly not tested every single edge of the TIPI emulation yet, so I expect to find some issues here and there. If MAME stops suddenly, this looks like a programming error that I have to fix. Did you try to enable logging with "-log" or "-oslog"?

[+9640News]

58 minutes ago, mizapf said:

I'd need some more detailed descriptions to test it. I have certainly not tested every single edge of the TIPI emulation yet, so I expect to find some issues here and there. If MAME stops suddenly, this looks like a programming error that I have to fix. Did you try to enable logging with "-log" or "-oslog"?

OK, here is what I got with -log

 

 

 

error.log

[+jedimatt42]

The TIPI TCP stuff is just messages using the same TIPI ROM functions that the TIPI DSRs use to send messages to and from for disk access. 

[+mizapf]

MID interrupt; this happens inside the emulation: An invalid opcode was encountered. As it seems, the page that is mapped to execution page 0 is empty (at least at its beginning). Maybe you should use the debugger, set a breakpoint at 0068, and check the mapper at F110. The initial cause is not shown here, but the return vector should be there.

[+9640News]

I used MAME's debugger and set a breakpoint at 0068 and MAME still crashed where I was not able to see what was at >F110.

 

I will send you privately a set of files for you to review.

 

Beery

[Eric Lafortune]

For a new project, I'd like to create a cartridge file with a paged 8 MB ROM. I've started with layout <pcb type="paged377">, but I'm running into its 2 MB limit. I'm now considering a multicart setup with 4 such cartridges, but that seems a bit cumbersome to manage. Are there any other options? Sorry if I'm missing something obvious.

[+mizapf]

Can a 74LS377-paged cartridge manage more than 256 pages * 8 KiB = 2048 KiB?

[+Lee Stewart]

1 hour ago, mizapf said:

Can a 74LS377-paged cartridge manage more than 256 pages * 8 KiB = 2048 KiB?

 

With only 8 inputs/outputs, I would think not.

 

Perhaps 2 chips could be ganged for more pages?

 

...lee

[+dhe]

@mizapf would you mind giving me the command line for a pure TI pbox and one for an ALL SNUG System?

 

As far as you know, is there any problems with sharing a scsi hard drive image between an ALL SNUG system and Geneve Based system (or course not accessing at the same time)?

[+mizapf]

@Eric LafortuneIf you tell me how you plan to build such a cartridge, I'd be ready to add that to the set of cartridge types. The point is that I cannot just extend the 377 type, not only because it does not reflect what you can do with a 377, but I simply don't know how to map so many pages.

[+mizapf]

@dheFrom this list (https://www.mizapf.de/en/ti99/mame/emufeat) what do you want to see in the boxes?

 

Currently I don't have the SNUG ASCSI in the list, so you would have to pick the WHTech SCSI. I don't see a problem with using the same SCSI drive in one or the other system.

[Asmusr]

30 minutes ago, mizapf said:

but I simply don't know how to map so many pages

Just extend the existing system with more bits controlled by writing to more words in cartridge space. That's how the cart Dragons Lair cartridge works (the demo at least). Classic99 and JS99er supports it.

DragonsLairDemo.rpk

Edited May 4, 2022 by Asmusr

[+mizapf]

I just checked my implementation; I would not suggest that you use the paged378 but gromemu instead. @Eric LafortuneThis should be the proper way for you, too. I planned the type gromemu as a catch-all for all kinds of GROM emulations (by EPROM) and also for the widest range of ROM sizes. It measures the ROM size and creates a bank mask from it. So, for setting the bank, in the extreme case you can write from 6000 to 7ffe (even addresses), i.e. up to 4096 banks by 8 KiB and hence up to 32 MiB.

[+dhe]

@mizapf you really need a buy me a coffee on your webpage! ?

 

 

Original TI:

  TI-99/4A

  Peripheral Expansion Box

      32K Memory Expansion Card

      RS232/PIO card (TI)

      TI Floppy Disk Controller Card

                  DSK1 - DS/SD   <= Which can be made with the TI Image Tool.

                  DSK2 - DS/SD

 

      Questions:

            Is it possible to 'print' from the PIO port?

 

SNUG Configuration:

   SGCPU (SNUG, TI-99/4A in the box)

    Peripheral Expansion Box     

        RS232/PIO card (TI)

        Enhanced Video Processor Card 1 (SNUG)   

        High Speed GPL Card (SNUG)

        BwG Floppy Disk Controller (SNUG)

                  DSK1 - DS/DD

                  DSK2 - DS/DD

 

        Two Items I see missing from the SNUG portfolio -

              High Speed RAM Disk - which is pretty kewl.

              SNUG Enhanced Speech Card.

 

Reference:

https://www.mizapf.de/en/ti99/mame/emufeat

[Eric Lafortune]

Thanks @mizapf -- gromemu is exactly what I need! It didn't occur to me that its ROM is paged too. Building new hardware would be beyond my capabilities.

[Eric Lafortune]

Actually trying it out now... The gromemu type seems to be subject to the same safety check as paged378: "Fatal error: Cartridge ROM size exceeding 2 MiB" (cartridges.cpp, line 148, introduced in Mame 0.205). I guess the check is not intended for gromemu?