*BSD News Article 12379


Return to BSD News archive

Newsgroups: comp.os.386bsd.questions
Path: sserve!manuel.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!pacbell.com!att-out!cbnewsj!dwex
From: dwex@cbnewsj.cb.att.com (david.e.wexelblat)
Subject: Re: How to select 72MHz on ET4000 board?
Organization: AT&T
Date: Fri, 5 Mar 1993 17:21:57 GMT
Message-ID: <1993Mar5.172157.6332@cbnewsj.cb.att.com>
References: <1993Feb25.151244.15114@gmd.de> <rcpt.730720661@rwa.urc.tue.nl> <1993Mar4.171002.5808@super.org>
Lines: 77

In article <1993Mar4.171002.5808@super.org> becker@super.org (Donald J. Becker) writes:
> In article <rcpt.730720661@rwa.urc.tue.nl> rcpt@urc.tue.nl writes:
> >veit@borneo.gmd.de (Holger Veit) writes:
> >>In article <rcpt.730581552@rwb.urc.tue.nl>, rcpt@rwb.urc.tue.nl (Piet Tutelaers) writes:
> >>|> 
> >>|> The ET4000 board in our 486DX33 box does provide a 72MHz clock in
> >>|> order to achieve a 70 Hz refresh rate for 1024x768 mode. The manual
> >>|> explains how to select this frequency with VMODE.EXE an MSDOS utility
> >>|> (mode 0x38).
> >>|> 
> >>|> If I startup `startx' and write its output into a file I don't see this
> >>|> 72MHz clock. Does this mean I can't select this clock with XFree1.2?
> >
> >>The clock might be select with a mechanism that is not in the area
> >>of the standard 16 clocks. This is not supported by XFree86. Another
> >>possibility is that your VMODE (or the handbook) lies about the 72MHz;
> >>also already seen. There is a facility to run an external clock
> >>program (if you manage to find out how the 72MHz clock is enabled), 
> >>see the XFree86-1.2 doc on this.
> 
> Xfree86 doesn't usually detect the correct clocks in my ET4000.  It
> has some clocks that are far abover the 65Mhz-nominal-max rating of
> the ET4000, but it Xfree often screws up the lower-rate clocks as
> well.  You can either
> 
> o run Xfree a bunch of times on an idle machine, and guess at which
>   values are correct, or
> o Add a 'Clocks' line that just numbers the clocks 0 to 15, and add
>   'modes' and resolution lines that use every clock.  Step through the
>   clock rates to find one that you like.
> 
> Clocks 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
> Modes "1024x768-0" "1024x768-1" "1024x768-2" "1024x768-3" "1024x768-4"...
> 
>   If you are curious about the true dot-clock rates, measure the
>   horizontal sync with a frequncy counter and multiply away...
> 
> 				DB
> -- 
> Donald Becker				       becker@super.org
> Supercomputing Research Center
> 17100 Science Drive, Bowie MD 21114		   301-805-7482

XFree86 1.2 has improved the accuracy of the clock probing.  Have you
tried again with 1.2?

Many boards have 80MHz clocks.  I think that is is part of the problems
that some users see with snow/smearing on XFree86.  The spec for the
ET4000 is 65MHz, but the board is supplying higher clocks.

The reason that you see the REALLY high numbers (100+Mhz) is the divide-by-2
bit.  These clocks really do exist, but the BIOS doesn't use them (nor
should you).  The BIOS sets the ET4000's divide-by-2 bit (hence getting
a 65Mhz clock from a supplied 130MHz clock).  This makes it much easier
to generate clocks from 25Mhz to 80MHz - just use the divide-by-two when
you need to.

XFree86 detects 16 clocks for an ET4000, even though it only has 3 clock
select leads.  The reason is that we use the divide-by-two bit as an
"extra" clock select.

The other thing that I just discovered on the WD chipsets, and will
investigate for Tseng as well, is that some chipsets allow you to feed
MCLK as VCLK (i.e. use the memory-refresh as the video dot-clock).  Which
means that there are really more clocks available than XFree86 is detecting.
I will be looking at this for all of the chipset for which I have documentation
(not all of them, unfortunately), and adding support for this feature to
XFree86 1.2.1.  Anyhow, this may explain where the 72MHz clock is coming
from.

--
David Wexelblat <dwex@mtgzfs3.att.com>  (908) 957-5871
AT&T Bell Laboratories, 200 Laurel Ave - 3F-428, Middletown, NJ  07748

"In and around the lake, mountains come out of the sky.  They stand there."
	-- Yes, Roundabout