X68000 core

[frankmonk]

X68k core porting for MiSTer is completed.
Please check:
http://fpga8801.seesaa.net/archives/20200510-1.html

These are wonderful news! Thank you for your work on this

[Piacenti]

X68k core porting for MiSTer is completed.
Please check:
http://fpga8801.seesaa.net/archives/20200510-1.html

One billion times thank you!!!

[elexor]

I compiled a test version here
here's a video of it running some games
https://youtu.be/R8s2h4pdDzM

courtesy of rysha.

[friendlyjoe]

Puu様、
本当にありがとうございました。

Thank you so much.

[pstriolo]

Hello

Thank you for the great works. Can someone build the .rbf file so we can test it ?

Regards.

[Peredonov]

Amazing news!
Thanks to dshadoff for his help and of course most of all a huge thank you to puu for porting this core.

puuさん、お疲れ様！

[Sorgelig]

X68k core porting for MiSTer is completed.
Please check:
http://fpga8801.seesaa.net/archives/20200510-1.html

Can you release the source for NIOS firmware? Later i would like to check how to modify it for MiSTer API, so secondary SD card won't be required.

[uigiflip]

tried this beta core runs less than sorgelig's updated framework x68000 for me, for example super sf2 turbo loads on old core but not new, and many hdf when try to load state out of memory

[puu]

The software of NiosII is the same as that for DE0-CV. If you are in a hurry, see below.
https://fpga8801.up.seesaa.net/rtl/rele ... 200329.qar
This corresponds to diskemu.v2/software/nios-src.zip included in the archive.

[c64tone]

tried this beta core runs less than sorgelig's updated framework x68000 for me, for example super sf2 turbo loads on old core but not new, and many hdf when try to load state out of memory

Try running 'switch' from the command line and setting the memory to 4mb?

[Sorgelig]

The software of NiosII is the same as that for DE0-CV. If you are in a hurry, see below.
https://fpga8801.up.seesaa.net/rtl/rele ... 200329.qar
This corresponds to diskemu.v2/software/nios-src.zip included in the archive.

Thank you! I'm not in hurry, still busy by many other things.. Just though it would be good to have FW sources for paired core. I remember you've already released the FW sources before but it was unclear which HDL is compatible.
Ideally, would be good to be able to mount FDD/HDD from MiSTer OSD by choosing it from primary SD card.

[rondc]

@c64tone You're right. Using "switch" from Human prompt, you can select the RAM , from 1 to 12 MB.
It works !
SASI hd images works fine, so with bigger images for SCSI don't-

@Sorgelig
It's a very good plan. Many people gets losted when trying to configure and operate the core.

Regards

[puu]

The original X68000 supports SASI only. SCSI was added later. Therefore, it uses a tricky method to share the same DMA channel, which is not easy to handle.
With the X68000, even SASI supports 16 connections, so a total of 640MB can be used. I think that it is sufficient as the capacity at that time.

[puu]

Since the X68000's SASI port is just GPIO, it is possible to connect a SCSI device to the SASI port using the dedicated driver SXSI.
Since the SASI HDD emulation is performed by NiosII, it may be possible to easily use the SCSI image by rewriting this as SCSI emulation and operating SXSI in the X68000.

[friendlyjoe]

> Mr. Puu,

This is wonderful information. Thank you.
Finally we can appreciate X68000 again, hopefully very soon!

[Sorgelig]

The original X68000 supports SASI only. SCSI was added later. Therefore, it uses a tricky method to share the same DMA channel, which is not easy to handle.
With the X68000, even SASI supports 16 connections, so a total of 640MB can be used. I think that it is sufficient as the capacity at that time.

With ability to swap images, 16 connections aren't required. I think up to 4 is enough. Especially when it will be available through OSD.

[mrbig6251]

I like to read that kind of good news , congratulations guyz


Mr BIG

[LFT]

Thank you puu for your work.
I am having a lot of fun exploring the X68000 core.

[softtest1]

Thank you Puu for this core and anyone who has helped with porting. I can't wait to try it out!

With ability to swap images, 16 connections aren't required. I think up to 4 is enough. Especially when it will be available through OSD.

Would the images be hotswappable (no reset)? The X68000 has quite a broad software library, including operating systems, emulators, music and painting applications. So I think that putting as much data as possible on one drive might provide a better workflow for some users.

With that said, swapping would still be very useful for gaming.

[puu]

SASI HDD images are swappable, realtime. But I don't know it works because O/S(human) wasn't designed for hot-swapping.

[Sorgelig]

SASI HDD images are swappable, realtime. But I don't know it works because O/S(human) wasn't designed for hot-swapping.

no problem to reboot after changing the HDD.
Btw, are you planning to add support FDD the same way through mounted files?

[Grabulosaure]

SASI HDD images [...]

.

Hello.

I've made a preliminary version of the X68000 core that directly uses MiSTer framework for HDD (instead of the NIOS software and second SD card).
Select the disk image then "Hard Disk : MiSTer"

http://temlib.org/pub/mister/release-F68kMiSTer.rbf
http://temlib.org/pub/mister/F68K.tar.gz.

I have adapted an SCSI-to-MiSTer disk emulation module to the SASI interface and this X68k core (from SS5/20). Coding is... peculiar as it is based on microcode, but it supports many SCSI commands, and can be extended (source is scsi_mist.vhs, the script asm.mist.rb generates scsi_mist.vhd ).

- Code isn't optimised, transfers could be faster, but it's already faster than the direct SD interface (the main issue is properly handling 256 bytes sectors)
- Only one hard disk implemented. It can be extended to several units either by replicating the block, or some internal logic (this SCSI disk never "disconnects", so interleaved transfers between several units don't need to be handled.)
- Tested with SASI ".HDF" files. I don't know how SCSI disks are interfaced, some signals are different (for example the SCSI ATN signal). The .HDF files are raw dumps of disk contents, ".HDS" files for SCSI disk images may use a different format, I don't know.
- Of course, the NIOS system also handle floppies and SRAM, which should be adapted as well.

[Sorgelig]

SASI HDD images [...]

.

Hello.

I've made a preliminary version of the X68000 core that directly uses MiSTer framework for HDD (instead of the NIOS software and second SD card).
Select the disk image then "Hard Disk : MiSTer"

http://temlib.org/pub/mister/release-F68kMiSTer.rbf
http://temlib.org/pub/mister/F68K.tar.gz.

I have adapted an SCSI-to-MiSTer disk emulation module to the SASI interface and this X68k core (from SS5/20). Coding is... peculiar as it is based on microcode, but it supports many SCSI commands, and can be extended (source is scsi_mist.vhs, the script asm.mist.rb generates scsi_mist.vhd ).

- Code isn't optimised, transfers could be faster, but it's already faster than the direct SD interface (the main issue is properly handling 256 bytes sectors)
- Only one hard disk implemented. It can be extended to several units either by replicating the block, or some internal logic (this SCSI disk never "disconnects", so interleaved transfers between several units don't need to be handled.)
- Tested with SASI ".HDF" files. I don't know how SCSI disks are interfaced, some signals are different (for example the SCSI ATN signal). The .HDF files are raw dumps of disk contents, ".HDS" files for SCSI disk images may use a different format, I don't know.
- Of course, the NIOS system also handle floppies and SRAM, which should be adapted as well.

Good job! I like the idea of NIOS-less core.
MiSTer supports up to 4 mounted images at the same time. SRAM probably needs read/write mode, so it will consume one mounted slot. Probably 2xHDD + 1xFDD + SRAM should be an optimal config.
Another way is to move SCSI/SASI/FDD layers to HPS side like it's done in Minimig and some other cores - in this case there won't be limit of 4 images.