Danny Milosavljevic schrieb: > Hi, > > I would like to extend directfb to support Portrait Mode (rotated TFT screen).
Nice, that's really a missing feature! > I found some earlier posts, but seems it stalled. > > Anyway, attached is a tiny extension (not very complete and untested, just to > give you an idea what I'm up to) to the nvidia driver to support rotating an > image (for now by 270 degrees). We should be able to use the TMU for that. > I also had a look at src/core/layer_region.c:dfb_layer_region_flip_update, > but I don't have enough knowledge yet to "connect it all together" from here. > > If DirectFB always uses a back buffer, the obvious and least-impact way would > be to have the back buffer(s) be for example 768x1024 and rotate-update the > front buffer to 1024x768. Right, dfb_layer_region_flip_update() is the central point where rotation should happen. However, adding support for differently sized buffers in one surface will be difficult. > However, most Flippers seem to just set some offset register on the graphics > adapter, so that isn't really possible. The window stack updates are blitted from back to front buffer most of the time, i.e. when only a portion of the screen is affected. Fullscreen applications/updates could be forced to do a blit for flipping. > My earlier tries included swapping every x and y in nvidia_2d.c, which worked > kind of, but I guess DirectFB mostly works over the frame buffer and so not > all things were actually rotated. > > What do you think? I think limiting the rotation support to blitted flipping is fine. PDAs for example would use that mode anyways, because of the windowing. -- Best regards, Denis Oliver Kropp .------------------------------------------. | DirectFB - Hardware accelerated graphics | | http://www.directfb.org/ | "------------------------------------------" _______________________________________________ directfb-dev mailing list [email protected] http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
