Subject: Re: framebuffers
To: None <ccallana@vision.elee.calpoly.edu>
From: Gordon W. Ross <gwr@mc.com>
List: port-sun3
Date: 02/23/1996 12:44:07
> From: Chris Callanan <ccallana@vision.elee.calpoly.edu>
> Date: Fri, 23 Feb 1996 09:26:43 -0800 (PST)

> > It is fairly easy to add support for new memory-mapped display
> > boards.  One just needs to figure out how X expects the memory
> > map for the device to be laid out (for mmap access).
> > 
> > Getting probe right is harder.  (I still don't have the probe
> > quite right for some of the frame-buffer drivers...)
> 
> what happens if probe is wrong???

The hard part with probe is NOT recognizing some other type of
video board as the one your driver supports.  The problem is,
most video boards look like a memory board (all look the same).
One has to kludge around looking for minor variations between
boards, and disallow "impossible" configurations, etc.

If the drivers all attach (even those that should not) then
you could end up opening mismatched driver/hardware and get
rather strange results as (i.e.) the cg4 driver tries to talk
to a cg6 board...  (So, don't open the "false match" ones! 8^)

Gordon