Re: amigappc test kernel for CSPPC and BPPC

[Frank Wille <frank%phoenix.owl.de@localhost>]

Andreas Barth wrote:

> http://u1.ipernity.com/17/85/74/9628574.a30d4a3f.jpg
>
> kernel boots, drives configuration is the same, but now with a medium
> in the CD-ROM drive, looks not ok, ends up in debug console. On Amiga
> side, the CD-ROM works ok.

There are known problems with DMA on amigappc. Here the hardware was found
to be in an unknown state after an interrupt. When you recompile a kernel
without DEBUG and without DDB in the config file it will continue running,
but probably without success. Maybe afsc.c also needs some membarriers,
which I already inserted into ahsc.c.

Better would be to edit sys/arch/amiga/dev/siop.c and change line 131 to
int siop_no_dma = 1;

On my A3000 the same driver works ok without DMA.


> http://u1.ipernity.com/17/85/75/9628575.27a2bbaa.jpg
>
> tried to use the debug console, most things don't seem to work. Should
> it?

Yes. Functions like ps and show map do work here. Maybe something was
trashed?

Does ddb work when you don't insert a CD and boot with "askroot", then enter
ddb manually by typing "ddb" for the root device?


> The CV643D shows garbage, when I try to boot netbsd from an fully
> booted Amiga system, means also the graphics card is already in use. It
> only works so far, when I run netbsd before graphics card was
> initialized by Amiga drivers. Yes, the background is black.

I have never seen a CV643D. How does it behave under NetBSD/amiga (68k)?
Should be the same.

The current gobsdppc-solution is just temporary. When amigappc ever becomes
a usable port we would install a boot-block for booting, as with amiga68k.
Then this problem doesn't matter.

-- 
Frank Wille