New MAME release

[+dhe]

So would the command line be:

    mame genmod -bios 1.00  -peb:slot8 idec -peb:slot8:ide:h1 generic -peb:slot8:ide:h2 generic -peb:slot8:ide:h3 -peb:slot8:ide:h4 -hard1 ide1.HD -hard2 ide2.HD -hard3 ide3.HD -hard4 ide4.HD  

 

    What tool do I use to prepare a blank ide harddrive/partition image?

 

[+dhe]

Looks like TI Image Tool is the way to go for the IDE partitions.

 

I think originally I used chdman to create HFDC images.

[+mizapf]

Yes, TIMT3 will help you set up an IDE drive with partitions. You can then put SYSTEM-SYS and LOAD-IDE on the partition, maybe also a suitable AUTOEXEC file. You should, however, use bios 2.00 so that you can boot from it.

 

The command line looks like this:

 

mame geneve -bios 2.00 ... -peb:slot7 ide -peb:slot7:ide:ata:0 hdd -hard1 myidedrive.hd -peb:slot8 ddcc1 $*

 

The connector is "ata", and you can connect up to two drives (at ata:0 and ata:1, as -hard1 and -hard2). There is no "generic" or other option, as those are specific to the HFDC and MFM drives.

 

Note that you can set up up to 8 partitions on the first drive. Each partition will be seen as a separate drive (IDE1...IDE8). So you don't have to connect multiple drives.

[+dhe]

Thanks, that should be all the pieces that my brain was missing to get me going!

[+dhe]

PSA - listdevices is awesome.

 

D:\mame-ide>mame geneve -bios 2.00 -window -oslog -peb:slot7 ide -peb:slot7:ide:ata:0 hdd -hard1 myidedrive.hd -peb:slot8 ddcc1  -flop1 dsdd1.dsk -listdevices
Driver geneve (Geneve 9640):
   <root>                         Geneve 9640
     colorbus                     V9938 Color bus
     dram                         RAM
     gatearray                    Geneve Gate Array
     joyport                      TI-99 Joystick port
       twinjoy                    TI-99/4(A) Twin Joysticks
     kbd                          PC keyboard connector
       geneve_kb_101              Geneve XT Keyboard 101 Keys (HLE)
     maincpu                      Texas Instruments TMS9995 @ 12.00 MHz
     mm58274c                     National Semiconductor MM58274C RTC
     pal                          Geneve PAL circuit
     peb                          Peripheral expansion box Geneve
       slot2                      TI P-Box slot
       slot3                      TI P-Box slot
       slot4                      TI P-Box slot
       slot5                      TI P-Box slot
       slot6                      TI P-Box slot
       slot7                      TI P-Box slot
         ide                      Nouspikel IDE interface card
           ata                    ATA Interface
             0                    ATA Connector
               hdd                IDE Hard Disk
                 image            Harddisk
             1                    ATA Connector
           atalatch_even          Octal Registered Transceiver
           atalatch_odd           Octal Registered Transceiver
           bq4842                 Benchmarq BQ4842 RTC
           bq4847                 Benchmarq BQ4847 RTC
           bq4852                 Benchmarq BQ4852 RTC
           crulatch               74LS259 Addressable Latch
           rtc65271               Epson RTC-65271 RTC
           sram512                Buffered SRAM
       slot8                      TI P-Box slot
         ddcc1                    Myarc Disk Controller Card
           0                      Floppy drive connector abstraction
             525dd                5.25" double density floppy drive
               floppysound        Floppy sound @ 44.10 kHz
               flopsndout         Speaker
           1                      Floppy drive connector abstraction
             525dd                5.25" double density floppy drive
               floppysound        Floppy sound @ 44.10 kHz
               flopsndout         Speaker
           2                      Floppy drive connector abstraction
           3                      Floppy drive connector abstraction
           ls259                  74LS259 Addressable Latch
           pal                    Myarc DDCC-1 PAL u1
           ram                    RAM
           wd1770                 Western Digital WD1770 FDC @ 8.00 MHz
           wd1772                 Western Digital WD1772 FDC @ 8.00 MHz
     pfm512                       ATMEL 29C040 512Kx8 FEEPROM
     pfm512a                      ATMEL 29C040A 512Kx8 FEEPROM
     screen                       Video Screen @ 21.47 MHz
     sound_out                    Speaker
     soundchip                    SN76496 @ 3.57 MHz
     sram                         RAM
     sramexp                      RAM
     tms9901                      TMS9901 Programmable System Interface
     vdp                          Yamaha V9938 VDP @ 21.47 MHz

D:\mame-ide>

 

[+dhe]

Troubles, Maybe a toggle or a flag I don't understand?

 

  Startup: mame geneve -bios 2.00 -window -oslog -peb:slot7 ide -peb:slot7:ide:ata:0 hdd -hard1 myidedrive.hd -peb:slot8 ddcc1  -flop1 dsdd.dsk

 

 

 

 

 

 

 

[+mizapf]

I have "Map at boot time: SRAM". The DIP setting "RTC" was primarily used for the RTC-65271 and Thierry's RTC ROM boot loader.

[+mizapf]

6 hours ago, dhe said:

PSA - listdevices is awesome.

There are some more interesting options:

 

-showusage: shows all command line options

-listslots: list available slots and slot devices (here: connectors and connectable devices)

-listbios: list BIOS options for the system

-listmedia: list available media for the system

-listsoftware: list known software for the system (as XML file, basically outputting the ti99_cart.xml)

[+dhe]

 

error.log

[+mizapf]

I was successful with loading the DSR into the IDE card on a TI-99/4A. Could you probably try that first? In my test, the BQ4847 was properly detected (using DSR 1.6).

 

For the Geneve, you may have to use ROMPAGE in GPL mode; I'll test that later.

 

 

[+dhe]

When I updated the DSR on the physical Geneve, I loaded TIMODE in the autoexec and ran exec ideload - with the dsr files on dsk1.

The card was automatically seen at cru >1900.

 

I think there is something off,  because on Mame, it doesn't see the card automatically at >1900.

 

Ref:

  

For my testing of gdm2k, I don't believe I need fine grain control over the exact configuration - clock, ram, partition size. I just want to load up a bunch of different types of files and directory levels X deep and try copies and moves.

 

Mame will provide the perfect test bed, as I will be able to backup the starting configuration with a batch file, and play until (if) a breakage occurs, recording exactly what steps I had taken.

 

FWIW - TI Image Tool V3 has been used extensively to create floppy images, the IDE drive for the upgrade and testing both on the real geneve and the Mame emulated Geneve and has worked flawlessly.

 

Thank you for the work on both Mame and TI Image Tool.

[+mizapf]

In most cases, when you change DIP switch or Machine configuration settings, you have to restart the emulation. There are some devices (e.g. GRAM Kracker) that are notified of switch changes, but not all, and in particular not the IDE card.

 

Also, you must set the "Genmod decoding: off" when using a Geneve. Only use it for the Genmod, and then don't forget to also set "Genmod TI mode: on", or you won't be able to turn on TIMODE.

 

See the attached disk image. You can boot the Geneve from it, then type &NEWIDE at the MDOS prompt. It works for me. You can also run &STDEA which loads GPL mode and the Editor/Assembler cartridge.

 

@InsaneMultitasker@9640News

When I stop IDELOAD in EXEC (2.7), I'm getting an undefined opcode in the log. I'm not sure which one is causing it. Also, there is a message "Unsupported GPL access attempted, aborting", sometimes in graphics mode.

 

[:maincpu] Undefined opcode 0000 at logical address 0024, will trigger MID

 

This looks like a BLWP @>0000 (with the usual vector WP=83E0, PC=0024) while there is nothing at that location.

newide.dsk

[+dhe]

>In most cases, when you change DIP switch or Machine configuration settings, you have to restart the emulation. There are some devices (e.g. GRAM Kracker) that are notified of switch changes, but not all, and in particular not the IDE card.

 

Thanks, I did not know this. From now on, I will simple make the changes, drop to the command prompt, restart and verify the changes "took".

 

>Also, you must set the "Genmod decoding: off" when using a Geneve. Only use it for the Genmod, and then don't forget to also set "Genmod TI mode: on", or you won't be able to turn on TIMODE.

 

Ok. That makes sense. When Genmod is enacted, it automatically gives you 2MB, correct? Which would mean decoding would need to be on. On regular hardware Genmod, if your memory card is 512K AMX decoding is not necessary?

 

Thanks again for the corrections in my understanding, I will give this another try this weekend.  If I fail, I should probably just try geneve and only add genmod, when I've got things going.

 

[+InsaneMultitasker]

4 hours ago, mizapf said:

@InsaneMultitasker@9640News

When I stop IDELOAD in EXEC (2.7), I'm getting an undefined opcode in the log. I'm not sure which one is causing it. Also, there is a message "Unsupported GPL access attempted, aborting", sometimes in graphics mode.

 

[:maincpu] Undefined opcode 0000 at logical address 0024, will trigger MID

 

Hmmm.  I thought that you identified this MID some time ago and that it was fixed.  EXEC 2.8 is the latest release, so we should try that version with DU2K.   It is most likely a memory mapper issue - either in EXEC or in DU2K. If I recall correctly, Fred handles page >BA mapping in his DM2K/DU2K programs. 

 

With respect to the "unsupported GPL access" message: keep in mind that EXEC only emulates a small number of GPL routines; the GROMs are not loaded*.  If a program attempts to use one of the many GROM routines that are not emulated, EXEC will terminate.  I thought Fred steered clear of GPL routines so maybe this issue is caused by the MID.  I suggest trying the program with the TIMODE interpreter (GPL).   You can pursue an OS/EXEC issue in the Geneve OS dev topic, if it comes to that.

 

*If you invoke TIMODE (or TIMODE2) and execute GPL, then return to the command line and execute EXEC, the GROM and ROM code survive in the TIMODE reserved pages.  Long ago, I considered clearing the GROM pages upon exit (from GPL) but I'm not sure that would solve anything.

 

[+mizapf]

@F.G. KaalWould you mind sharing the source code of IDELOAD with me so that I can more easily find a good location for a breakpoint?

[+mizapf]

Same with EXEC28 (running IDELOAD). I'm running GeneveOS 7.42.

[+dhe]

@mizapf with all of your help I was able to partition the drives (TI Image Maker V3), install the DSR (having all the flags set correctly) and Initialize the paritions. w00t!

 

 

 

 

Then I loaded SCSMAP 1A, ASSIGN E=SCS1:

 COPIED IN ORDER LOAD-IDE, SYSTEM-SYS.

 

System failed to boot!

 

 

 

 

@9640News sent me a prototype LOAD-IDE that is multicru aware, I replaced the standard issue LOAD-IDE, with the multicru aware version, and boot it did.

 

 

 

I think something is keeping MDOS from detecting the card, even thou in the menu's it's set to CRU >1900 ?

 

Now, I'm trying to remember why I did all of this. 😃

Edited June 27 by dhe
move pics to correct places.

[+mizapf]

Isn't it "LOAD-IDE"? On your partition you have a "LOADIDE".

[+dhe]

Sorry photo out of order.

 

 

Boot with Multi-CRU Aware load-ide. - No Boot with standard load-ide.

[+mizapf]

So it does load with LOAD-IDE 3.00 - that sounds good enough, doesn't it?

[+InsaneMultitasker]

On 6/24/2024 at 5:02 PM, mizapf said:

@InsaneMultitasker@9640News

When I stop IDELOAD in EXEC (2.7), I'm getting an undefined opcode in the log. I'm not sure which one is causing it. Also, there is a message "Unsupported GPL access attempted, aborting", sometimes in graphics mode.

 

[:maincpu] Undefined opcode 0000 at logical address 0024, will trigger MID

 

This looks like a BLW

The MID issue has been fixed.  I'll post an updated EXEC in the Geneve topic.  As you are aware, thanks to your testing and our discussion and sleuthing, c99 exits by returning to ROM address >0024. This is the /4a RESET vector address.  Unfortunately, EXEC's reset vector is different.  The fix was to add code to >0024 that branches to the right reset address.  

[+dhe]

2 hours ago, InsaneMultitasker said:

The MID issue has been fixed

What does MID stand for?

[+mizapf]

Macro Instruction Detect

 

According to the datasheet, it is intended to be used for new opcodes that are not implemented by microcode inside the CPU but by a machine language program in memory.

 

We've been talking about it several times, see e.g.

 

 

[+9640News]

@dhe

 

Dan, the only thing the LOAD-IDE did differently was search for a CRU instead of using default >1900 CRU, and then separately, it was either the SCSI or IDE version that did not set the appropriate AUTOEXEC boot tracking.

 

I believe your initial earlier LOAD-IDE issue was the result of using a hard drive image incompatible with the IDE image.  It wasn't until you used TIIMAGETOOL that got it straightened out.

 

At least from the posts I was able to follow, I suspect that was the issue. 

 

Beery

[+dhe]

12 minutes ago, 9640News said:

I believe your initial earlier LOAD-IDE issue was the result of using a hard drive image incompatible with the IDE image.  It wasn't until you used TIIMAGETOOL that got it straightened out.

Incorrect. When building an IDE MAME config, I started with a HD I built with TIMTV3. It was finally setting the correct Flags in MAME config that allowed me to copy files to the hd/part.

But, it still wouldn't boot from the IDE with the original LOAD-IDE, IE. The IDE Controller or LOAD-IDE didn't work, even though the controller itself, was configured at CRU >1900. Using the updated LOAD-IDE you sent by PM, allowed me then to boot off the IDE controller.

 

The original problem I had with GDM2K and copy move was on real hardware, which is why I built this image so I could more easily try to 'trip up' gdm2k during the move cycle.