Re: CSG5710

Hello Bild,

sorry for the late answer:

* On Thu, May 02, 2013 at 03:57:01PM -0400 Bil Herd wrote:
 
> So if you're saying someone changed the memory map when cost reducing we
> tended to castrate someone that screwed with compatibility.

Yes, they did:

Have a look into http://cbmrom.trikaliotis.net/listings/1571-05.lst
(regular 1571) and http://cbmrom.trikaliotis.net/listings/1571cr-01.lst
(1571 CR from 128DCR).

It starts at $883C. The code at $883C-$886B is completely missing from
the CR. But, the author did not try to cover this by adding NOPs -
instead, he just removed the code completely!

From L88F0 on ($88F0 on 1571, $88C0 on 1571CR) it gets even worse.

Or have a look into L89B3 ($89B3 on 1571, $8991 on 1571CR) and see how
there are seemingly "minor" differences between both variants (with seem
to be because of the way the 5710 is programmed as opposed to the way
the WD177x is programmed). These differences add to even more
discrepancy - the code addresses from both firmware variants are moving
away more and more... At $8FA4, both variants meet again. However,
everything inside is not compatible.

Thus, any code that tries to access MFM disks by using the firmware
directly will break. However, this is not the only place. At $AA7C, both
variants start to differ again - by ONE BYTE only! Have a look at LAAB6,
which is at $AAB6 on the 1571, but on $AAB5 on the 1571CR. Again, this
difference makes much parts from there on unusable.

Taking into account that $ADD8 - $BEFF is completely unused on the
1571CR, I cannot believe these patches were not achievable differently.


I believe the distinction 1571CR to regular 1571 took me more time to
combine into one source (with ifdefs) than all the other variants (2031,
1540, 1541-01, 02, -03, -05, -06, 1570, 1571-01 and -03) TOGETHER!

Regards,
Spiro.

-- 
Spiro R. Trikaliotis
http://www.trikaliotis.net/

       Message was sent through the cbm-hackers mailing list