Path: utzoo!utgpu!attcan!uunet!cbmvax!grr
From: grr@cbmvax.UUCP (George Robbins)
Newsgroups: comp.sys.amiga.tech
Subject: Re: Hopalong
Message-ID: <4644@cbmvax.UUCP>
Date: 2 Sep 88 21:01:00 GMT
References: <2581@sugar.uu.net> <4957@netnews.upenn.edu> <7004@well.UUCP>
Reply-To: grr@cbmvax.UUCP (George Robbins)
Organization: Commodore Technology, West Chester, PA
Lines: 46

In article <7004@well.UUCP> ewhac@well.UUCP (Leo 'Bols Ewhac' Schwab) writes:
> 
> 	Which brings me to the reason I ported this over to c.s.a.t:  I'd
> like to bounce off you people what I think is wrong with my setup.  I have
> an ASDG Mr. C with 2M of RAM in it, and a Ronin Hurricane card 14.foo MHz
> '020, 16 MHz '881.
> 
> 	When the ASDG cage is disconnected, everything runs great;
> 'Hopalong', as well as all other programs, run without incident.  When I
> plug in the cage, vanilla programs run fine.  However, programs utilizing
> the '881 are prone to random crashes, typically exception code 4 (illegal
> instruction), 11 (line F emulator), and to a lesser extent 10 (line A
> emulator), as well as one instance of 14 (format error).  Precisely when
> these crashes occur is random.
> 
> 	I suspect this is due to a combination of the kickstart tower PAL's,
> and a flakey '881.  If it were just the PAL's, then one would expect ordinary
> code running on the '020 to bomb randomly as well (which it once did before
> I had the kickstart tower PAL's grounded).  If, on the other hand, it were
> just the '881, then one would expect the random crashes to persist without
> the cage attached.  Ergo:  It's both the PAL's and the '881 that need
> looking at.

Well, it's hard to tell without looking at the hardware involved, but
I suspect that the Hurricane card isn't doing quite enough kludgery
to make everything work in your expanded configuration.

The Amiga hardware generally doesn't decode the FC (function code) lines
emitted by the processor - it just sees a single flat 16Mbyte address
space.  With a 68020 various function codes are emitted during the
coprocessor interface protocol that *do not* correspond to normal
external memory cycles.  This can cause a conflict if a memory card
reponds or bus buffers are enabled.

The general rule is that the 68020 adapter logic must gate the address
strobe signal so that only "memory" and "interrupt acknowledge" cycles
are visble to the Amiga and expansion bus. 

> 	Is my theory correct?  Is it close?  Am I from Neptune on this one?

no, maybe, perhaps...

-- 
George Robbins - now working for,	uucp: {uunet|ihnp4|rutgers}!cbmvax!grr
but no way officially representing	arpa: cbmvax!grr@uunet.uu.net
Commodore, Engineering Department	fone: 215-431-9255 (only by moonlite)