On Thu, 2004-12-16 at 15:17 -0500, Branden Robinson wrote: > > What precisely does disabling the "DRI" module in the XF86Config-4 > accomplish that making the DRI modules unavailable by removing the > xlibmesa-dri package doesn't?
It prevents the X server from initializing and using the DRM and the DRI in general. > If I had to guess, I'd say that codepaths usually found in XFree86's > driver/$FOO/FOO_dri.c don't get followed. That's one example, yes. > And in that case, any bug which doesn't go away when xlibmesa-dri is > removed but does when the "DRI" module is disables properly belongs in > the xserver-xfree86 package. Or the kernel. > 1) try to repro with LIBGL_ALWAYS_INDIRECT=1 > 2) try to repro with xlibmesa-dri removed 2) is kinda redundant, both have the same effect, the only difference being that 1) doesn't even try to load the DRI driver whereas 2) tries but fails. It may be a good idea to set LIBGL_DEBUG=verbose as well though to make sure that LIBGL_ALWAYS_INDIRECT=1 is actually in effect. Beware that if it is, there will _not_ be any special output, whereas if it's not, there will be something like libGL: XF86DRIGetClientDriverName: 4.0.1 r200 (screen 0) libGL: OpenDriver: trying /home/michdaen/src/mesa-cvs/Mesa/lib/r200_dri.so drmOpenByBusid: Searching for BusID pci:0000:00:10.0 drmOpenDevice: node name is /dev/dri/card0 drmOpenDevice: open result is 4, (OK) drmOpenByBusid: drmOpenMinor returns 4 drmOpenByBusid: drmGetBusid reports pci:0000:00:10.0 ... > 3) try to repro with 'Load "DRI"' commented out of XF86Config-4 That's a good second step. -- Earthling Michel Dänzer | Debian (powerpc), X and DRI developer Libre software enthusiast | http://svcs.affero.net/rm.php?r=daenzer
