[NOTE: this goes to both amiga and port-amiga lists]

Steve wrote:
>  I have been struggling with this problem for over a year now and
>  really haven't had much success since NetBSD 1.1 (mainly due to
>  the fact I have no known, tested "working" grfmode file.. 
>  Anyone have one?? =)

Here's what I have been using:

#
# Monitor definition file used by grfconfig
#
#num clock wid  hgt dep hbs  hss  hse  hbe  ht   vbs vss vse vbe vt
1 65000000 1024 768 8   1024 1088 1232 1244 1344 768 785 791 806 806
2 31500000 640  480 8   640  712  752  832  832  480 500 503 525 525
3 42610000 800  600 8   840  896  960  1040 1040 600 630 640 670 670
4 42610000 800  600 15  792  920  1040 1080 1080 599 630 640 670 670
5 42610000 800  600 16  792  920  1040 1080 1080 599 630 640 670 670
#6 31500000 640  480 24  640  712  752  832  832  480 500 503 525 525
c 42610000 800  600 4   840  896  960  1040 1040 600 630 640 670 670

In addition to the console mode I use mode 1 - I'm not sure the other
modes even work :)

>  Anyone who is also working with the '040+A2000+Picasso, please
>  drop me a line!  I have a feeling the fix is gonna be something
>  real simple, just not readily apparent!

For Amiga 2000 + Zeus + PicassoII it suffices to access VDAC_DATA in a
funny address:

[.. snip from grf_clreg.h ..]
/* Video DAC */
#define VDAC_ADDRESS		0x03c8
#define VDAC_ADDRESS_W		0x03c8
#define VDAC_ADDRESS_R		0x03c7
#define VDAC_STATE		0x03c7
#ifdef M68040
extern int cltype;
#define VDAC_DATA		((cltype == PICASSO) ? \
				 (0x03c9 + 0xfff) : 0x03c9)
#else
#define VDAC_DATA		0x03c9
#endif
#define VDAC_MASK		0x03c6
[.. snip from grf_clreg.h ..]

I've been using the above for quite some time, but as Steve has told
me it is not enough for GFORCE. And the above is not good enough for a
general patch as for example A3000+040 should not be patched..

I tried to ask Klaus Burkert for information about the 040 problem but
so far I've got no answer from him.