Matrox G200 mga_vid_test problem

[Paul Novak]

I'm having trouble getting mga_vid to work on my system. I tried to follow
the installation instruction exactly and have read the FAQs, documentation
and searched google and haven't figured it out yet.

I'm providing all the information I can think of, which is more than
most people will wan to see...

I have an old Matrox Milleninium AGP G200 that came with 8MB of memory.
Afterwards (many years ago), I bought the 8MB memory expension card
and installed it in the graphics card to give it 16MB of memory.

I am running Slackware 10.1 without the gnome and kde packages installed.
I have compiled my own 2.4.29 kernel.

Xorg -version reports:

X Window System Version 6.8.1

The device section of my /etc/X11/xorg.conf file looks like this:

Section "Device"
Identifier "Card0"
Driver "mga"
VendorName "Matrox"
BoardName "MGA G200 AGP"
BusID "PCI:1:0:0"
Option "NoHal" "true"
Option "DRI" "true"
Videoram 16384
EndSection

I do know that if I do not have the "Videoram" line in my xorg.conf file,
when I start up X, my Xorg.0.log will show that X detects 8MB of memory
in the video card instead of 16.

cat /proc/mtrr shows this:

reg00: base=0x00000000 ( 0MB), size= 512MB: write-back, count=1
reg01: base=0x20000000 ( 512MB), size= 256MB: write-back, count=1
reg02: base=0xdb000000 (3504MB), size= 16MB: write-combining, count=2
reg05: base=0xd0000000 (3328MB), size= 128MB: write-combining, count=2

I guess reg02 is my Matrox card.

I downloaded and installed the MPlayer-1.0pre6a source, cd-ed to drivers,
and ran make. mga_vid.o and mga_vid_test both compiled. /dev/mga_vid
already existed (I did not run mknode to create it, it must have come
created in Slackware 10.1). ls -l /dev/mga_vid shows:

crw-rw-rw- 1 root root 178, 0 2001-03-25 15:04 /dev/mga_vid

I ran "insmod mga_vid.o", and checked with dmesg and saw this:
"mga_vid: detected RAMSIZE is 8 MB", so I ran "rmmod mga_vid" and
then "insmod mga_vid.o mga_ram_size=16". dmesg then shows:

Matrox MGA G200/G400/G450/G550 YUV Video interface v2.01 (c) Aaron Holtzman & A'rpi
mga_vid: Found MGA G200 AGP at 01:00.0 [Matrox Graphics, Inc. MGA G200 AGP]
mga_vid: MMIO at 0xf2196000 IRQ: 9 framebuffer: 0xDB000000
mga_vid: OPTION word: 0x4049CD21 mem: 0x13 SGRAM
mga_vid: RAMSIZE forced to 16 MB
syncfb (mga): IRQ disabled in mga_vid.c
mga_vid: 1 supported cards found

I then ran mga_vid_test. Test 1 (green square) is fine. When test 2 (cool mosaic
pattern) is starts, I do not see a "mosaic pattern": I see the results of test 3
(color blend). When test 3 starts, the display does not change: I still see
the color blend.

Just to see what would happen, I removed mga_vid and re-instered it with
only 8MB of memory and verified the result with dmesg. I reran mga_vid_test
with identical results. I didn't continue to compile mplayer because I think
I have something wrong.

Should all 3 mga_vid_test tests work for me? If yes, what did I do wrong and
what can I do to fix it?

Paul.

[Attila Kinali]

On Wed, 13 Apr 2005 18:11:51 -0500

Post by Paul Novak
I have an old Matrox Milleninium AGP G200 that came with 8MB of memory.
Afterwards (many years ago), I bought the 8MB memory expension card
and installed it in the graphics card to give it 16MB of memory.

Ok

Post by Paul Novak
I am running Slackware 10.1 without the gnome and kde packages installed.
I have compiled my own 2.4.29 kernel.

Ok
mga_vid is totaly unrelated to X. It actualy may interfere with it
in some cases as both directly access the graphics card.

Post by Paul Novak
"mga_vid: detected RAMSIZE is 8 MB", so I ran "rmmod mga_vid" and
Matrox MGA G200/G400/G450/G550 YUV Video interface v2.01 (c) Aaron Holtzman & A'rpi
mga_vid: Found MGA G200 AGP at 01:00.0 [Matrox Graphics, Inc. MGA G200 AGP]
mga_vid: MMIO at 0xf2196000 IRQ: 9 framebuffer: 0xDB000000
mga_vid: OPTION word: 0x4049CD21 mem: 0x13 SGRAM
mga_vid: RAMSIZE forced to 16 MB
syncfb (mga): IRQ disabled in mga_vid.c
mga_vid: 1 supported cards found

That looks ok.

Post by Paul Novak
I then ran mga_vid_test. Test 1 (green square) is fine. When test 2 (cool mosaic
pattern) is starts, I do not see a "mosaic pattern": I see the results of test 3
(color blend). When test 3 starts, the display does not change: I still see
the color blend.

I do not quite get what you see. Could you elaborate it a little bit ?

And could you also tell me whether you run mga_vid_test on a (framebuffer)
console or in X ?

Post by Paul Novak
Just to see what would happen, I removed mga_vid and re-instered it with
only 8MB of memory and verified the result with dmesg. I reran mga_vid_test
with identical results. I didn't continue to compile mplayer because I think
I have something wrong.

Try it anyways. It might be a non-error.

Post by Paul Novak
Should all 3 mga_vid_test tests work for me? If yes, what did I do wrong and
what can I do to fix it?

Actualy they should. Though because i lack a PCI G200 i wasn't able
to test the code for it for quite a long time.


Attila Kinali

[Paul Novak]

Post by Attila Kinali
On Wed, 13 Apr 2005 18:11:51 -0500

Post by Paul Novak
I then ran mga_vid_test. Test 1 (green square) is fine. When test 2 (cool mosaic
pattern) is starts, I do not see a "mosaic pattern": I see the results of test 3
(color blend). When test 3 starts, the display does not change: I still see
the color blend.

I do not quite get what you see. Could you elaborate it a little bit ?
And could you also tell me whether you run mga_vid_test on a (framebuffer)
console or in X ?

I ran the test X. I'd try it on a console window, but I don't want to do that
until tomorrow because I'm seeding a bittorrent and I don't want to kill it
until tomorrow.

Test 1 runs fine: I see the green square. Then test 2 says it is starting,
saying I should see a mosaic pattern. I do not see the mosaid pattern: instead
I see a color blend. When test 3 starts, saying I should see a color blend,
I continue to see the color blend; the display does not change. In other words,
in my case, the results of test 2 are exactly the same as the results of test 3.

Post by Attila Kinali

Post by Paul Novak
Just to see what would happen, I removed mga_vid and re-instered it with
only 8MB of memory and verified the result with dmesg. I reran mga_vid_test
with identical results. I didn't continue to compile mplayer because I think
I have something wrong.

Try it anyways. It might be a non-error.

I did try it and unfortunatly it did not work for me. (In X), when I use -vo xmga,
the picture I see isn't viewable: there's a lot of horizontal lines across the
window. Without -vo xmga, using whatever mplayer defaults to, the picture is very
good quality. Makes me wonder why I wan to get xmga to work at all :-)

Post by Attila Kinali

Post by Paul Novak
Should all 3 mga_vid_test tests work for me? If yes, what did I do wrong and
what can I do to fix it?

Actualy they should. Though because i lack a PCI G200 i wasn't able
to test the code for it for quite a long time.
Attila Kinali

My card is a G200 AGP, not a PCI; I don't know if that matters.

Thank you kindly for your suggestions. Any additional help would be greatly
appreciated.

Paul Novak
pdnovak (at) comcast (dot) net