On Tue, 2003-07-29 at 12:17, Soeren Sonnenburg wrote: > On Tue, 2003-07-29 at 02:32, Michel Dänzer wrote: > > On Mon, 2003-07-28 at 18:42, Soeren Sonnenburg wrote: > > > On Mon, 2003-07-28 at 14:30, Michel Dänzer wrote: > > > > On Mon, 2003-07-28 at 14:02, Soeren Sonnenburg wrote: > > > [...] > > > > > > No no, not the alternating black/white pixels, but a 99.99% white > > > > > > screen > > > > > > with some black dots. > > > > > > > > > > ahh sorry, external display... internal is fine. > > > > > > > > I suspect this is a bug triggered by the explicit Modes "1280x854". If > > > > it works without, I have an idea how to fix it for the next version. > > > > > > You were right it works without... I am looking forward for your next > > > (fixed) version :-) > > > > What do you need the Modes "1280x854" for though? > > Hmmhhh... I really intensively tested now all the possible > configurations (single/dualhead with and without xinerama) and must say > wow! it is working very stably... I saw one problem once though, that > when I switch from X to console and used less to view some log file and > scrolled backwards, there was a stripe reproducing itself... > Unfortunately I cannot reproduce this... and I was unable to trigger it > again...
The radeonfb console acceleration can get confused when running X without Option "UseFBDev". I use for i in off on; do fbset -accel $i; done as a workaround. But maybe it even works with "UseFBDev"? The behaviour is undefined, but that doesn't mean we can't be lucky. :) > > Yes, the driver probes for an external display via DDC and disables > > output if it doesn't detect one. You could override this with Option > > "MonitorLayout" though. > > True... Indeed if I leave out the above Modes option and start X without > an external CRT but later on attach one, I also get some output on the > monitor... it is a bit darker in the top though (wrong mode...)... > > With attached monitor the resolution of the external monitor is choosen, > which is why I felt like I need the Mode option... Maybe infact I don't. You may indeed need to limit the virtual resolution, and your method would have worked if it wasn't for the bug. > I am still looking for 'the' solution, where one uses the internal DFP > most of the time then connects an external CRT/DFP and makes use of that > one without having to restart X... > > I thought I could do this by using a multihead config with two displays > :0.0 and :0.1 but there I had the problem that the mouse cursor was > somehow only on :0.1 and I did not know how to set the mouse to be on > :0.0 ... The mouse pointer should work the same way as with Xinerama, but moving windows between the heads probably requires special support as found e.g. in GTK as of 2.2. > A different option would be to use a xinerama mode, and have a program > to somehow en/disable the mouse going to the external display... Unfortunately, I don't know of a way to achieve this, but again I suspect that the mouse would be less of a problem than windows. > Very good job Michael ! But I only do my small part... thank all the people working hard on XFree86, in particular Hui Yu for his great radeon driver work. > It is a pity that XFree86 is not yet capable of adding displays > dynamically like osx can... that would indeed "rock". True. The combination of http://bugs.xfree86.org/show_bug.cgi?id=276 and xrandr might be interesting. -- Earthling Michel Dänzer \ Debian (powerpc), XFree86 and DRI developer Software libre enthusiast \ http://svcs.affero.net/rm.php?r=daenzer