I have two bugs open on the mga driver that I'd like some feedback on:

http://bugs.xfree86.org/show_bug.cgi?id=1098

This one implements I2C support for G-series cards with single-chip
dualhead support. (i.e. not G200 MMS)  The purpose is to talk to the
MGA-TVO (Maven) chip which controls everything related to TV output, but
is also used for one of two monitors in a dualhead setup. There is no
public documentation for Maven, so I based my code on matroxfb and
OpenBeOS code.  (None was directly used.)  The end result is that DPMS
and DDC both work on the Maven head, and also that we can tell whether a
monitor is attached or not (if not, we can set timings appropriate for a
TV).  We also should be able to detect the version of the Maven chip,
which will let us get G200-TV working, which uses an earlier version --
I implemented code for that detection based on matroxfb but I'm not sure
if it is working correctly until I find a G200-TV of my own.

I wish I knew how DVI support worked with it.  AFAIK, no G400 was sold
with a DVI port, but there exists an MAFC add-on:
http://www.matrox.com/mga/products/upgrades/flat_panel_g400.cfm
It looks like it can be added onto even dualhead cards as long as the
upgrade connector is not occupied.  I guess that is the reason they
claim that it doesn't work on a single-head G400 that has been upgraded
to dualhead.

This is the first steps towards getting rid of HALlib completely.  It
will not longer be necessary once we are doing our own mode setup on all
heads of all G400 types, because G400 is the last one who needs the HALlib
still.  The TV mode setup is easy because Petr has done most of the
reverse engineering already.  I would really like to find someone that
can check out the DVI angle (or send me a DVI monitor -- on loan of
course!)

Does anyone know what the second port on a G400-TV is:
http://www.tomshardware.com/video/19991118/images/matrox_pcb.gif
It would appear to be a DVI port from that side angle, but it must be
for some special break-out cable with a VGA and NTSC plugs on it.


http://bugs.xfree86.org/show_bug.cgi?id=1401

This one is a quickie fix for a crash which occurs under the described
conditions.  Without this, it is impossible to run a static X server
with a G400 and a dualhead configuration, or to run a dualhead
configuration without HAL and have a graceful failure -- as soon as any
display update occurs, you get a server segfault.

-- 
Ryan Underwood, <[EMAIL PROTECTED]>

Attachment: signature.asc
Description: Digital signature

Reply via email to