Samsung SyncMaster 172T DVI

[elanthis]

Hi all,

I can't seem to get the Samsung SyncMaster 172T monitor working over DVI in X. The monitor uses DVI great during boot up and in text mode; however, if I try to start X (using the 3123 version drivers - I can't use the 4xxx drivers because of the Render bug and my all GNOME/GTK2 desktop), the monitor changes resolution a few times (like normal), then goes into what looks like sleep/standby mode (the green light turns orange, screen is all black). A few seconds later, the Capslock and Scrolllock keyboard leds start blinking in unison. I can't get back to a terminal or ctrl-alt-del or ctrl-alt-backspace; I think the machine is actually locking up (I don't have a net connection handy to test otherwise), and I have to power-cycle to get it back.

The nv driver doesn't seem to work any better.

This happens if I set the ConnectedMonitor to DFP. Leaving that off, it seems to work fine (using Analog mode). Using the Debian get-edid command reports that EDID isn't fully supported by the monitor; that might explain it (another post on the forums said EDID has to work for the drivers to use DVI mode). However, the XFree86.0.log claims that EDID does work. Totally confused. ^,^

This is all using a Geforce3Ti 500, on Debian/Sid (XFree86 4.2 w/ patches).

Attacking my XF86Config-4 file.

[bwkaz]

Originally posted by elanthis
... A few seconds later, the Capslock and Scrolllock keyboard leds start blinking in unison.

That would be a kernel panic or a kernel oops or something similar.

There may be something in /var/log/messages about the oops/panic, but I don't know; it is possible that klogd might have not been able to grab it.

Which kernel are you running? Depending on the oops, it might be originating with marginal hardware (memory especially -- maybe run memtest86 overnight sometime?), or there might just be bugs in your kernel. Though bugs are probably not likely in Sid...

[elanthis]

kernel 2.4.19, self-compiled with some ext patches (acls, attrs), and other debian patches.

The logs just show the NVdriver loading up AGPGART, and after are the log entries from the next bootup.

I'll run memtest86 when I get home tonight from work. I don't think the memory is bad, I tend to use all of it and then some and never have applications crash or misbehave, but it's still a possibility I suppose.

[elanthis]

I ran the memtest86, there are no problems.

[ToastyX]

To anyone that might run into this thread while searching, see my reply here: http://www.nvnews.net/vbulletin/show...279#post278279