ACA630 + NetBSD = success
Radosław Kujawa
Just a quick report: The new ACA630 turbo board for Amiga 600, made by Individual Computers is fully compatible with NetBSD. This card features 68030 25MHz and 32MB of Fast RAM, which is more than enough to run basic installation. Amiga 600 equipped with this hardware is a nice, little, NetBSD-capable machine (though a bit slow - installation over network took almost 3h, and that was without X11).
Btw. A600 is misdetected as A1200, because bootblock identifies A1200 by looking for card.resource. What do you think about identifying A1200 by reading Alice chip revision, and then looking for card.resource?
Frank Wille
Rados?aw Kujawa <radosla...@c0ff33.net> wrote:
> Just a quick report: The new ACA630 turbo board for Amiga 600, made
> by Individual Computers is fully compatible with NetBSD. This card
> features 68030 25MHz and 32MB of Fast RAM, which is more than enough
> to run basic installation.
Nice. I expected nothing else. :)
Similarly interesting is that the FPU emulation works. I have never seen
that before, although I knew it exists.
> Btw. A600 is misdetected as A1200, because bootblock identifies A1200
> by looking for card.resource.
I doubt that NetBSD ran on many A600s before.
> What do you think about identifying
> A1200 by reading Alice chip revision, and then looking for
> card.resource?
Yes, that should identify an A600 safely. You can make a patch if you
like.
You would also have to adapt autoconf.c (new is_a600() function, removing
the comments in pccard detection) and machdep.c (print model name).
--
Frank Wille
--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-...@muc.de
Ignatios Souvatzis
> On Tue, 11 Jan 2011 10:10:49 +0100
> Rados?aw Kujawa <radosla...@c0ff33.net> wrote:
>
> > by Individual Computers is fully compatible with NetBSD. This card
> > features 68030 25MHz and 32MB of Fast RAM, which is more than enough
> > to run basic installation.
>
>
> Similarly interesting is that the FPU emulation works. I have never seen
> that before, although I knew it exists.
Ha! I have in my now semi-functional A3000 a switch so that I can change
the mc68882 chipselect from chipselect to bus error (which signals to the
68030 that no FPU is installed, thus creating the F-line trap.
> I doubt that NetBSD ran on many A600s before.
But it did, once, *nearly* run on an A500. Unfortunately, the A590
... oh wait. If we had an A590, we didn't have the A530, or does
the A530 have a SCSI on-board? Anyway, we had problems with the
SCSI, probably because the wd33c93 on the board was even older than
the one in the A3000 and had its own set of ... wierd interpretation
of the programming manual[1], so our driver back then would hang
sometimes.
-is
[1] It would be impolite to call them bugs... I think mlelstv
mentioned, let's say, unexpected combinations of interupt *class*
and detailed reason code, while the reason code would still be
correct - is that right?
Radosław Kujawa
On Jan 11, 2011, at 2:13 PM, Ignatios Souvatzis wrote:
> On Tue, Jan 11, 2011 at 01:45:41PM +0100, Frank Wille wrote:
>> Similarly interesting is that the FPU emulation works. I have never seen
>> that before, although I knew it exists.
>
> Ha! I have in my now semi-functional A3000 a switch so that I can change
> the mc68882 chipselect from chipselect to bus error (which signals to the
> 68030 that no FPU is installed, thus creating the F-line trap.
Nice. I was somewhat surprised that FPU emulation works so well. There were some problems in the past. For example: http://mail-index.netbsd.org/port-amiga/2005/01/25/0007.html . I remember having problems with FPU emulation back in the 1.6.x days.
>> I doubt that NetBSD ran on many A600s before.
>
> But it did, once, *nearly* run on an A500. Unfortunately, the A590
> ... oh wait. If we had an A590, we didn't have the A530, or does
> the A530 have a SCSI on-board?
Yes, GVP A530 does have a on-board AM33C93A.
> Anyway, we had problems with the
> SCSI, probably because the wd33c93 on the board was even older than
> the one in the A3000
Yes, but I think it would work with A590, because it's very similar to A2091. We could have 68030 with A590 - some turbo boards installed in CPU socket were produced.
Another interesting fact about A530: it does have proprietary ASIC inside, designed by GVP and manufactured by Motorola. Not PLD but real ASIC. I wonder how many A530s were produced. I wish I had one ;). It's pretty easy to buy working A500, but almost impossible to find turbo board with MMU.
On Jan 11, 2011, at 1:45 PM, Frank Wille wrote:
>> What do you think about identifying
>> A1200 by reading Alice chip revision, and then looking for
>> card.resource?
>
> Yes, that should identify an A600 safely. You can make a patch if you
> like.
I'll experiment with it.
> You would also have to adapt autoconf.c (new is_a600() function, removing
> the comments in pccard detection) and machdep.c (print model name).
Ok, I knew about that ;).
--
Best regards,
Radosław Kujawa
Buecheler, Torsten
Somewhere in a paperboard in our basement, there is A4000 with a cyberstorm 060 50MHz and approx. 80MB Ram.
AFAIR it contains several additional cards (Ariadne, ArcNet, CyberVision64, Multiface, Toccata, SCSI_II Controler, VLab Motion, maybe more) and I exchanged some of the original chips (FatAgnus, SCSI Controller(?), ...)
The last time I run this machine was with Amiga OS 3.1 and NetBSD 1.3.1(!).
As you might imagine, I took good care of this machine. Anyway, I will not use it anymore. So, if there is somebody who will definitely run the machine, we surely can make a deal. Easiest would be if this one could fetch the machine, so I don't need to ship all this stuff. I am located near Speyer in Germany.
Best regards,
Torsten Bücheler
Ele...@vip.hr
I'm interested!
If we agree, she will be running and get the best care.
Please contact me: elerion(at)vip.hr
BR,
E.
Radosław Kujawa
On Jan 11, 2011, at 3:15 PM, Radosław Kujawa wrote:
> On Jan 11, 2011, at 1:45 PM, Frank Wille wrote:
>
>>> What do you think about identifying
>>> A1200 by reading Alice chip revision, and then looking for
>>> card.resource?
>>
>> Yes, that should identify an A600 safely. You can make a patch if you
>> like.
> I'll experiment with it.
Patch attached, tested on A600 and A1200. Additional tests on A1200s are welcome.
Frank Wille
> Patch attached, tested on A600 and A1200. Additional tests on A1200s
> are welcome.
Your boot.amiga and kernel modifications work on my A3000 too. Looks good.
I will commit your patch tomorrow. Thanks.
--
Frank Wille
Radosław Kujawa
On Jan 12, 2011, at 7:53 PM, Frank Wille wrote:
> Radoslaw Kujawa wrote:
>
>> Patch attached, tested on A600 and A1200. Additional tests on A1200s
>> are welcome.
>
> Your boot.amiga and kernel modifications work on my A3000 too. Looks good.
> I will commit your patch tomorrow. Thanks.
Great, please commit bppcsc too.
--
Best regards,
Radosław Kujawa
Frank Wille
>> Your boot.amiga and kernel modifications work on my A3000 too. Looks
>> good. I will commit your patch tomorrow. Thanks.
>
> Great, please commit bppcsc too.
Done.
Man pages are always a bit difficult (for me) to get is right. I needed some
release builds until I was sure I didn't break anything. ;)
For the man page I took your license from bppcsc.c. Apart from that I only
fixed a few typos and style issues.
Radosław Kujawa
On Jan 15, 2011, at 1:11 PM, Frank Wille wrote:
> Radoslaw Kujawa wrote:
>
>>> Your boot.amiga and kernel modifications work on my A3000 too. Looks
>>> good. I will commit your patch tomorrow. Thanks.
>>
>> Great, please commit bppcsc too.
>
> Man pages are always a bit difficult (for me) to get is right. I needed some
> release builds until I was sure I didn't break anything. ;)
>
> For the man page I took your license from bppcsc.c. Apart from that I only
> fixed a few typos and style issues.
Thanks!