Re: HP 9000 Series 332 support. Hardware request.

[David Brownlee <abs%netbsd.org@localhost>]

On 27 September 2013 22:00, Andrew Gillham <gillham%gmail.com@localhost> wrote:
>
> Ok, trying to read causes a trap, not just write.
> What seems to work is tc->regs. objects like fbid, frame buffer /
> display dimensions etc.
> What is broken is tc->{plane_mask,cindex,rdata,gdata,bdata,strobe} and
> 'tccm_waitbusy' as well.
>
> With a return at the top of topcat_setcolor() and the monitor hooked
> up, serial console disabled it comes right up.  The display seems to
> work fine in the minimal tests I did before switching back to serial
> console.
>
> Any ideas how to detect this limited topcat and not access
> non-existent registers?  I don't have an HP332 option or HP_332 model
> id at this point.  There is too much magic cookie stuff in locore.s
> for model number recognition for me to figure it out.
> I would think if tc->num_planes==1 we could skip the topcat_setcolor
> stuff but I don't really know.

If sc->sc_fb->planes == 1 is enough for _attach() to report it as a
HP98544 topcat then I would have thought it a sufficient check to
disable all the colour related logic as well. I'm assuming the various
topcats are fixed bitdepth, so a HP98547 could not run in 1bpp.
Certainly the driver assumes that :)

Possibly just by wrapping the for statements around the two calls to
topcat_setcolor() with an if (fb->planes > 1).