*BSD News Article 11320


Return to BSD News archive

Received: by minnie.vk1xwt.ampr.org with NNTP
	id AA1571 ; Tue, 23 Feb 93 14:48:59 EST
Path: sserve!manuel.anu.edu.au!munnari.oz.au!uunet!ukma!darwin.sura.net!mojo.eng.umd.edu!floyd
From: floyd@eng.umd.edu (Mike Tierney)
Newsgroups: comp.unix.bsd
Subject: Re: driver mods for Diamond 24x (WD90C31)
Date: 16 Feb 1993 04:47:08 GMT
Organization: University of Maryland
Lines: 27
Message-ID: <1lprkcINN9bh@mojo.eng.umd.edu>
References: <1lnb3oINNf0e@mojo.eng.umd.edu> <JTSILLA.93Feb15112312@damon.ccs.northeastern.edu>
NNTP-Posting-Host: pepsi.eng.umd.edu
Keywords: XFree86 Diamond 24x

   An external program could be written to set the Diamond dot clocks using
the new method in XFree86 1.2.  However, my code changes are necessary to
initialize the WD90C31 chip regardless of the dot clocks.  I have not looked
at the new method, but it might be possible to set these registers
there.  If that is the case, the program would be rather small.

   Again, these mods do NOT support setting the dot clocks.  They only allow
the Diamond 24x to be used after DOS or some other program has pre-initialized
the dot clocks.

   I know there is code to set the dot clocks.  However it is possible that
the numbers used to set the clocks are not consistent from one 24X board
production run to the next.  This could be a support nightmare.  We already
know that from the 24 to the 24X the numbers changed.  I suppose given the
legal problems one could release object code or binaries, but this is not
a good method.  Still, it might be the only recourse those of use with 24X's
have to buying another card or getting sued.

   Would it be worth it to have a binary distribution only for the program
to set the clocks if someone had code to do it?  I hope there is another way.

   I think Diamond is getting left in the dust performance wise.  Maybe if
they changed their minds they might sell a few more cards before they
disappear.  :-)

   Mike Tierney
   floyd@pepsi.eng.umd.edu