In article <9408091621.AA26777@saul2.u.washington.edu>, donn@u.washington.edu (Donn Cave) writes:
> The current kernel support for the Z3 went in back in June.  A month or
> two before, Markus Wild had put some server code out on eunet, in the
> ``experimental'' directory.  (And kernel modifications, including the
> driver code from Lutz Vieweg that is the basis for the current driver.)
> 
> That experimental server code compiled for me without any serious
> modifications.  I compiled it with X11R6.  It's not perfect, but the
> flaws are not serious, at least not so serious that I was motivated to
> fix them.  The only two that come to mind are a square black rectangle
> that appears on the screen when the mouse runs into the edge of the screen,

This certainly results from the usage of a 64-pixel-wide hardware-cursor.
When I wrote the low-level sources, I trusted the NCR-77C32BLT manual which
decribed the 64-pixel HWC feature, meanwhile, NCR has stated that this
feature is broken in some chip revisions. We should switch back to use
only 32-pixel wide cursor sprite...

> It's reasonably fast, and Lutz
> apparently has ideas about how it could be speeded up

Yes, there are really great chances for speedups:

- The blitter is used very seldom, yet. There are a lot of drawing
  primitives which could be done by the blitter, if someone implements
  the appropriate call-back functions.
  
- The blitter supports "text-blits", which could speed up the text
  rendering by magnitudes.

> , although at present
> I don't think he's working on it.

Yes, that's unluckily correct. The reasons: My NetBSD installation was
kind of corrupted, so the ordinary X11R5 sources didn't compile, nothing
but trouble with missing/erraenous includes etc. I wanted to wait for
NetBSD 1.0 to come out which could give me a good chance for a clean
re-installation from scratch. But now I've got a second problem: I need
to have both my A3000 and my new HP-workstation on the same SCSI-bus,
as I need to access my MO-drive from both computers. This works well
with AmigaDOS, but NetBSD instantly crashes in this configuration.

Maybe someday I'll be able to disconnect the A3000 from the SCSI bus,
thus allowing me to run NetBSD on it again, but I cannot say when this
may happen.

If anyone wants to work on the RZ3 server code, I'll gladly support him
with new low-level routines and information.

cu, Lutz Vieweg



------------------------------------------------------------------------------