Amix in WinUAE

[Toni Wilen]

[Moved from MMU thread]

Amix at least asks for tape now

EDIT: Actually it does not get that far, it tries to find SCSI HD but it accesses DMAC/SCSI chip quite differently compared to A590/A2091/A3000 Amiga boot roms and it hangs.. Need to implement missing WD SCSI chip functionality first.

[mark_k]

A little googling mentions a 200MB HD size. So it could be worth trying something smaller. Didn't older A590/2091 drivers fail with >1GB disks? Maybe there could be a similar bug in the AMIX code.

For partitioning, might you need to create a UNI\0 partition using HDToolBox first? (I don't know, is there a scanned version of the AMIX install guide online? I have a printed copy somewhere which I can try to find if you get stuck.)

Oh, I just found this. PDF manuals, HD image of installed AMIX OS, etc. etc.

Edit: See this post for a link to a PDF of the Installing Amiga UNIX manual.

[Toni Wilen]

Success! Problem was as simple as writing not working correctly.

Amix installer created correct partitions and finally it asks for tape

(winuae.zip updated)

[Ratte]

Allways nice to see WinUAE growing .. next feature tape-emulation with sequential datastream and forward/rewind function.
?!

btw.
HAPPY NEW YEAR!!!

[Toni Wilen]

I don't think I'll bother with SCSI tape drive emulation. Not worth the trouble just for Amix. At least 68030 MMU and WD SCSI emulation is used by other programs.

For example ATAPI-CD emulation is more important than Tape drive emulation

[mark_k]

You'd probably only need to implement a few commands/features for AMIX to install. But anyway...

(Maybe this should be separated into a separate "getting AMIX working in WinUAE thread"..?)

I downloaded the pre-installed AMIX hard disk image from http://amigaunix.com/. Apparently that was created on an A3000. It boots at least, but then immediately reports an error, see pic. Anyone got any ideas? It may just be a problem with my WinUAE config.

[Toni Wilen]

Set HD to SCSI6.

It boots, does some HD checks and then reboots.

[Toni Wilen]

Tiny fix in LINK m68k instruction bus fault handling (address register that LINK modifies was not saved = register had wrong data if this instruction had bus fault while writing to stack)

Pre-installed Amix works now!

[mark_k]

Great job! Here's a screenshot with X and Open Look running in A2024 mode.

[mark_k]

Don't know whether it's of any relevance, but AMIX seems to lock up if you have the Immediate blitter option enabled.

Also, if after logging in you run rn then press Ctrl-Z, this text is shown (the system doesn't crash):
NOTICE: User BUS ERROR at 1000, PC:1000 FAULT:6 PID:150 CMD:rn

[Toni Wilen]

Originally Posted by mark_k

Don't know whether it's of any relevance, but AMIX seems to lock up if you have the Immediate blitter option enabled.

Immediate blitter didn't check if blitter DMA was enabled (I think this has been broken long time ago) and blit was always finished immediately.

AMIX starts weird long blit from address zero but it never enables DMA, blit never starts. I am not sure whats the point, some kind of protection?

Also, if after logging in you run rn then press Ctrl-Z, this text is shown (the system doesn't crash):
NOTICE: User BUS ERROR at 1000, PC:1000 FAULT:6 PID:150 CMD:rn

This may be just a bug in the program, not handling CTRL-Z properly. Invalid MMU bus errors generally always crash the system.

[Toni Wilen]

Originally Posted by mark_k

Great job! Here's a screenshot with X and Open Look running in A2024 mode.

Totally useless but also really cool

[mark_k]

Originally Posted by Toni Wilen

Immediate blitter didn't check if blitter DMA was enabled (I think this has been broken long time ago) and blit was always finished immediately.

AMIX starts weird long blit from address zero but it never enables DMA, blit never starts. I am not sure whats the point, some kind of protection?

Does it try to set up/start that blit early-on in the boot sequence? Maybe the developer was trying to clear/initialise chip memory, but forgot to actually enable blitter DMA. And anything else which uses chip RAM clears it using the CPU anyway, so the problem was never noticed???

[neozeed]

Wow totally awesome!

I'm using the 2.6.0beta on win7 x64 and the image from amigaunix.com. However can't get it to boot

[mark_k]

There's probably something wrong with your WinUAE config. If you post it here I can take a look. Can you get it to work if you use the a3000_amix_test.uae config I posted above? (You'll need to change the SCSI ID of the hardfile to 6 before booting.)

[Gilloo]

useless things have no price and are awesome!

Here is the config. amix.uae

Hi guys, using the amix.uae config i have managed to boot into amix 1.1, from there i have used the boot and root img files from here http://amigaunix.com/tiki-list_file_...hp?galleryId=1, and dd'd them into adf files. I can upload them if anyone is interested. After booting these images i have been able to mount the amix 2.1 cpio into winuae and and start the install process. Its not quite done yet but its looking promissing.

Btw, to create the .adf images. I loaded amix 1.1, added a network interface and bound it to my ethernet adapter. This requires winpcap. once this is done I booted into amix and set up the interface, ie ifconfig en0 192.168.1.100 netmask 255.255.255.0, from there i ftp'd into my server and downloaded the boot.img and root_v2.img. I bunzip'd them first to save some time. then these images can be dd's to your virtual fdd, ie dd if=boot.img of=/dev/rdsk/fd0 this will take a few minutes. once boot is done you can swap out your virtual fdd (press f12) then load your root disk. this is done in the same manner, ie dd if=root_v2.img of=/dev/rdsk/fd0 again this will take a few minutes. once this is complete you can reboot your system and boot the first disk, there will then be a prompt to insert the root disk. A little caveat. Use RDB mode to set the drive geom on the cpio image. this should be set to scsi1. I have kept the target disk as scsi6.

It doesn't boot at all using the same config and amix image