Bug#512058: black screen on VGA port
I have a Thinkpad T43p with an ATI FireGL V3200 graphics card. The system
runs under etch (I upgraded from sarge some time ago), and I do an aptitude
update/upgrade around once a day. I do not use packages from etch-backports
or packages that I compiled myself.
In my xorg.conf, I used the radeon driver, which works fine for the display
of the laptop. Concerning the VGA port, there is however a strange behavior:
(1) Let's say I boot the machine such that X is not started, connect a
display or a projector and start X, then both the laptop display as well as
the external display show the same picture (I use MergedFB with clone and do
not use the option NoDCC). I can quit X and restart (having the external
display connected when X starts), and things work fine.
(2) Now I quit X, disconnect the external display, and start X again. Then
I connect the external display as a test, but as the frequencies are not
correct (strangely, the external display says that it receives a 1600x1200
picture, which is the natural resolution of the laptop display, independent
of the current resolution on the laptop display). The important point here
is not whether I actually connect the external display while X is running,
just that X starts without an external display connected. Now I quit X.
(3) I connect the external display and start X again. Then the strange
thing comes: I get the proper picture on the laptop display, I get an
external signal with the correct resolution and proper frequencies, but the
external display is ENTIRELY BLACK! I can do now whatever I like, but the
external display remains black. I can change the resolution and get black
displays in different resolutions, and so on. The only thing that helps now
is to reboot and get back to step (1).
I am pretty puzzled by this behavior, which is 100% reproducible. Looking
into the log files, there is no difference between restarting X as long as I
remain within step (1) (with the exception of the very first start when drm
is loaded). There is a difference between the log files produced in steps
(1) and (2), as the driver states that no DCC can be loaded. There is,
however, _no_ difference between log files in steps (1) and (3), so that the
driver seems not to realize that there is a difference in the external VGA
output. I do not include these log files here, but I can provide them if
needed, or any other information that might be helpful.
I understand that most people start X with a display connected, but for
laptop users the present behavior is rather strange. I could live with the
fact that I need to have the external display connected when X starts, but
that I have to reboot the machine after erroneously starting X without an
external display is pretty odd (I give a lot of talks with my laptop). I
would guess (as I have played around a lot with the xorg.conf file in the
last days, leaving the above behavior untouched) that the problem described
here is quite general. It is not influenced by adding CRT2HSync to xorg.conf
or similar measures.
Note that the problem does not appear when using the fglrx driver instead of
radeon. I tried that to see whether the graphic card is the issue or the
driver. I cannot use fglrx, however, as I would need to compile the kernel
module to get OpenGL hardware support (at least without the fglrx kernel
module, I do not have OpenGL hardware support).
> uname -a
Linux mupfel 2.6.24-etchnhalf.1-686 #1 SMP Tue Dec 2 07:56:33 UTC 2008 i686
> dpkg -s libc6 | grep ^Version