Hi,

one of the problems was caused by a bug in the file
"src/sys/arch/arm32/vidc/console/vidcconsole.c" in line 505:

old line (wrong):
                        if (vidc_currentmode->hder >= 800)
should be:
                        if (vidc_currentmode->hder >= 1024)

I do not know why (I do not know anything about the VIDC20) but so it
works for all my video modes.

But I found that the VIDC20 registers called VIDC_FSYNREG and VIDC_CONREG
seem to be programmed to contain other values in NetBSD and in RiscOS.
I think that because I tracked the above problem by setting the screen mode
from RiscOS and disabling all VIDC register writes from the kernel.

I could switch on all register writes in the kernel again without an effect,
only these both caused rubbish if applied alone, only if the kernel was
allowed to change both something normal was to be seen on the monitor.

If someone could shed some light on these strange things it would be very
nice.
 
-- 
Peter Teichmann

----------------------------------------------------------------------------
Email: teich-p@rcs.urz.tu-dresden.de  WWW: rcswww.urz.tu-dresden.de/~teich-p