🚀 go-pugleaf

I first noticed this running the X11R4 server (compiled under 4.0.3)
with a cgfour (now running 4.1.1).  On the cgfour, the default setup
of Xsun is for two screens - one an 8bit color, and the other a
monochrome screen, switched by the overlay plane.

The problem is this:

/dev/bwtwo0 is *supposed* to be the B&W screen of the cgfour (as I
understand it), and it *IS* what Xsun is writing to.  (Verified by
screendumping it and viewing the image - clients running on screen 1
appear in the image, as does the cursor.)

/dev/bwtwo1 is the screen that the console driver is writing to, and
the screen that appears when I slide the pointer off of Xsun's color
screen.  (Verified also by screendump.)

My first thought was that the kernel was somehow massively
misconfigured.  But the problem remains even with a generic kernel.
(And the bwtwo{0,1} configuration lines are the same as in 4.0.3.)

My next idea was that somehow the devices in /dev had gotten their
minor numbers messed up.  But they *DO* match what /dev/MAKEDEV
created them to be.

So, my questions:

  Is it supposed to work this way?  (Nah.)

  What the heck is wrong?  ("Must be hardware.")

Confused,
--
Brian
brsmith@cs.umn.edu