The file amicon-ecs.c.diff contains my modifications to amicon.c.

As you might have guessed the major difference is that amicon.c now
support ECS screenmodes. That means that you can now run Linux/68k in
the sames screen-modes as the AGA-machines, but with less colors (this
includes 640*480 noninterlaced and 800*600 interlaced). To take
advantage of these new screenmodes you need a 31kHz monitor or a
multisync monitor, if you want to be able to run all resolutions (just
like under AGA).

Besides the support for ECS I have added some new screenmodes, that
are similar to the screenmodes specified by Commodore under Kickstart
3.x. Therefore I suggest we remove the old screenmodes, as they are
"non-standard", and some of them does not work under ECS. The new
screenmodes are as follows:

o multiscan       640*480
o multiscan-lace  640*960
o super72         800*300 - 72Hz
o super72-lace    800*600 - 72Hz
o dblntsc         640*400 - 59Hz
o dblpal          640*512 - 52Hz
o euro72          640*400 - 71Hz

The modes are centered so they look good on my Acorn AKF 50 monitor,
but in the future I suggest we find a way of specifying an
(x,y)-offset for the bootstrap program, relative to the
screen-position. In this way everybody can ajust the modes to their
hardware.  As I have inserted these new modes, I'm not sure whether
the default screen-mode table is correct anymore (someone, please
check).

For those who understand, there are some important changes to the code
concerning the programming of the ECS/AGA hardware.

1. FMODE does not excist under ECS, and should therefore not be touch
on these machines.  

2. The old way of waiting for the PAL-line has been changed from
(diwstop_h-2) to (htotal-4) when the SHRES-bit is set. This was done
because (diwstop_h-2) didn't work under ECS in SHIRES. Why (htotal-4)
works I don't know, and therefore I would like someone with an AGA
computer to check these pacthes on his/her machine.

I suggest that the calculation of ddfstop is made dependend of FMODE,
as it changes with the value of FMODE.


Thats it for now - please check-out these patches and tell me if you
find/correct any problems caused by my code.  I can only say that
these patches should have been out long time ago, but I suddenly got a
reply from Carolyn Scheppner at Commodore (CATS) about 4 weeks ago,
and thefore waited for her next reply, in the hope of getting some
more facts on the PAL-line from her. Unfortunately I have received no
reply, and therefore decided to release the patches in their current
state.

Jes Sorensen 15/11-94
(jds@kom.auc.dk)

PS: Everything was done on my old A2000, so don't think you need an
A3000 or newer to get the nice screenmodes (just a Super-Denise) -
I run my Workbench in a 640*400 noninterlaced resolution.