Sam wrote:
>> The diffs are tacked onto the end of this message.  Basically, I commented
>> out references to the CG2 frame buffer.  It doesn't exist on the 4c, but
>> it may be needed for older Sun-4 machines, so what I did was the wrong
>> thing for any distribution.  I had a similar problem to the one that 
>> Dennis Ferguson reported with bw2reg.h.  The patched X11 R6 sources 
>> that I ended up with actually wanted a struct def from cg2reg.h.  There
>> is no cg2reg or cgtwo reg in the system (1 Aug snapshot) or sources
>> (16 Aug ) that I have.

I'm guessing that you commented out the references simply because you didn't
have /usr/include/machine/cgtworeg.h, because - as both you and Dennis wrote -
it will definitely be needed for any future efforts involving Sun-4's.

Dennis Ferguson responds:
>There is a sys/sparc/include/cgtworeg.h in the -current sources, date-stamped
>
> *      $Id: cgtworeg.h,v 1.1 1994/08/14 00:13:20 deraadt Exp $
>
>which needs to be present in /usr/include/machine for the CG2 support in the
>server to compile.  Your system is too old to have it in /usr/include, though
>the sources you have should have included it.

I wonder if this should be mentioned anywhere if X-related archives are being
kept ... I don't think "patch" has an option to do something like "Create this
new file from this `diff -rc /dev/null /usr/include/machine/cgtworeg.h' patch,
but not if there's already a file there".  So some kind of pointer/reminder
might be in order ...

>I'm not sure whether sun4's ever shipped with this hardware, but I knew that
>some sun3's did, and I thought I'd make sure that all the bits and pieces
>someone might need to build a server for sun3 hardware were there while I
>understood what those bits and pieces were.

Unless I'm forgetting something, *any* color VMEbus Sun-4 - a Sun-4/[1234]x0 of
any variety - has to use a framebuffer that's a CG2 or derivative; i.e. the CG5
was basically a double-buffered CG2 and supported by the CG2 driver; even the
CG9 or the hapless GP or GP/GB or GP2 or GP2/GB combos were degenerate-case
"supported" by the cgtwo0 driver.  Support for cgtwo0 (and bwtwo0, of course)
is central to a successful Sun-4 port.

	- Greg