I have tested this new version with my Samsung LD190G, but it still doesn't work properly. I tested the monitor on my laptop with Ubuntu 9.04. I can start X server on both screens (internal and samsung), but samsung monitor doesn't display anything. I can start "xterm" on both screens using ssh access tough, like; "xterm -display :0.0; xterm -display :0.1" works. If I start my laptop with Samsung monitor plugged, then nothing shows up on any monitors in the end. I started X server manually after I booted up the laptop without Samsung monitor in recovery mode. JAK <div class="gmail_quote">On Wed, Jul 8, 2009 at 1:35 AM, Roberto De Ioris <<a href="mailto:roberto@unbit.it">roberto@unbit.it</a>> wrote: <blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> Hi all, on <a href="http://projects.unbit.it/downloads/displaylink-mod-0.3.tar.gz" target="_blank">http://projects.unbit.it/downloads/displaylink-mod-0.3.tar.gz</a> you can download the first 'alpha' release of a totally refactored kernel module. This new branch includes dynamic mode settings and a prototype of a rock-solid framebuffer that can survives detach (and re-attach ?) of devices. The code has been split in different files, and some new ioctls has been added to support a new xorg driver (still working on it) with full xrandr support, included the merging of (virtually unlimited) displaylink devices in a unique (big) framebuffer. I will not work anymore on the udlfb driver, but i will release a 'stable' release of displaylink-mod as soon as i have solved all of the hotplug stuff. *** Some notes for tester *** - the driver allocates all the framebuffer space (1900*1200*2) so its unsuitable for embedded devices, i will add a module param to limit the amount of allocated memory (and so the resolution) - you can change resolution of fbcon using the fbset utility, the module setups the resolution always on the best supported mode. - when you detach a devices that is still in use, the framebuffer is not deallocated until the last process using it is killed. (but there is probably a race condition that lock the fbcon) - error handling (most of all on probing code) is horrible, so expect some oops in case of some allocation failure. As always, thanks to Unbit and Marvell for supporting my work. -- Roberto De Ioris <a href="http://unbit.it" target="_blank">http://unbit.it</a> JID: <a href="mailto:roberto@jabber.unbit.it">roberto@jabber.unbit.it</a> _______________________________________________ Libdlo mailing list <a href="mailto:Libdlo@lists.freedesktop.org">Libdlo@lists.freedesktop.org</a> <a href="http://lists.freedesktop.org/mailman/listinfo/libdlo" target="_blank">http://lists.freedesktop.org/mailman/listinfo/libdlo</a> </blockquote></div>