Hi, Has anyone had a Matrox G450 working using the mga linux driver? I'm trying to avoid having to build the Matrox driver, but things aren't looking good. Thanks, Jenny
On 5/4/05, Jenny Hopkins <hopkins.jenny@gmail.com> wrote:
Has anyone had a Matrox G450 working using the mga linux driver? I'm trying to avoid having to build the Matrox driver, but things aren't looking good.
I made great progress on this yesterday, upgraded to kernel 2.6, and finally got the two monitors working in a stretched mode in X (only as root, but I was working on it!). However, when I powered up the computer this morning, one of the monitors is not functioning. The back of the card has two sockets, one of which requires an adaptor from the second monitor. Whichever monitor is plugged into this socket has only the "No signal" message on the display. If I have only one monitor plugged in to first socket, I get a single console display as normal. If I plug only one monitor in to the second, adaptor socket, I can work blind and get some very odd behaviour. As the bios would display, the "No signal" message moves slowly down the screen then returns to the centre and is still at what would be the LILO menu waiting for input. If I hit enter, the "No signal" message moves slowly UP as I would see the kernel boot text scrolling on screen, then remaining still after machine has booted. I switched it off with both monitors working yesterday, and haven't touched or joggled the cables since. I'm utterly baffled. Can anyone help? Thanks, Jenny
On Thursday 05 May 2005 10:16 am, Jenny Hopkins wrote:
If I have only one monitor plugged in to first socket, I get a single console display as normal. If I plug only one monitor in to the second, adaptor socket, I can work blind and get some very odd behaviour. As the bios would display, the "No signal" message moves slowly down the screen then returns to the centre and is still at what would be the LILO menu waiting for input. If I hit enter, the "No signal" message moves slowly UP as I would see the kernel boot text scrolling on screen, then remaining still after machine has booted.
Pretty odd, but I would suggest that it is coincidence. Many screens move the "No Signal" banner around to avoid screen burn, if you unplug the monitor from the machine you should see a cycle of the same behaviour every few minutes. It is possible for some sync pulses getting through the VGA input to be effecting screen position, but usually the presence of these pulses would cause the "no signal" message to vanish. As to why it has stopped working. I can only think that it is some X configuration post startup that hadn't come into effect until you restarted the system....I assume screen two only fires up once X has started ? In very rare cases I have seen Monitors get confused and only accept a signal after they have been unplugged for a few moments. Wayne
On 5/5/05, Wayne Stallwood <ALUGlist@digimatic.plus.com> wrote:
As to why it has stopped working. I can only think that it is some X configuration post startup that hadn't come into effect until you restarted the system....I assume screen two only fires up once X has started ?
In very rare cases I have seen Monitors get confused and only accept a signal after they have been unplugged for a few moments.
Wayne: I wrangled with this all yesterday, with help from Adam. What seems to be happening is that my working XF86Config for sharing the image across the two monitors only shows half that display on one monitor. However, after starting X with a config that SHOULDN'T work, but does work and gives a mirrored two screen image, I can then put the first config back in place, restart X and get the setup I want with my display stretched across two monitors. (Thereafter both monitors work in console mode as well, but not before). We're pretty puzzled, because there doesn't seem to be anything being loaded/stated in the second config that isn't in the first. It's all a bit technical for me :-( Thanks, Jenny
On Friday 06 May 2005 12:02 pm, Jenny Hopkins wrote:
However, after starting X with a config that SHOULDN'T work, but does work and gives a mirrored two screen image, I can then put the first config back in place, restart X and get the setup I want with my display stretched across two monitors. (Thereafter both monitors work in console mode as well, but not before).
We're pretty puzzled, because there doesn't seem to be anything being loaded/stated in the second config that isn't in the first.
It's like there is something in the first config that initialises the card correctly (or in your case the 2nd output of the card) which is missing in the 2nd config...personally I thought that card initialisation was a BIOS job but maybe you have to do it for different resolutions or something. I had a similar thing with a soundcard a few years ago, it would only work after a cold boot if I booted Windows first and then rebooted into Linux.
On Fri, May 06, 2005 at 05:57:27PM +0100, Wayne Stallwood wrote:
On Friday 06 May 2005 12:02 pm, Jenny Hopkins wrote:
We're pretty puzzled, because there doesn't seem to be anything being loaded/stated in the second config that isn't in the first.
It's like there is something in the first config that initialises the card correctly (or in your case the 2nd output of the card) which is missing in the 2nd config...personally I thought that card initialisation was a BIOS job but maybe you have to do it for different resolutions or something.
What is the mad part is that the first config *shouldn't* work, and causes XFree86 to segfault (according to the logs) the config I conjured up by hand which should work doesn't until the first config has been run and segfaulted XFree86, in that segfault obviously something happens to the card which makes it run later on. After 3 hours of trying to make this work I got a bit of a headache and we gave up for the time being, at least she has a workaround for now to make the dual head work :) I think this situation would be perhaps best investigated at a kit meet in the future where multiple observers can look for weird behaviour. Adam -- jabberid = quinophex@jabber.earth.li AFFS || http://www.affs.org.uk/ || Not a filesystem newbaby!! http://quinophex.org/newbaby/
participants (3)
-
Adam Bower -
Jenny Hopkins -
Wayne Stallwood