On 23/01/15 19:24, Haixia Shi wrote: > Hi Emil, > > On Fri, Jan 23, 2015 at 8:42 AM, Emil Velikov <emil.l.veli...@gmail.com> > wrote: >> Might be worth having a look at how platform_drm does it. But we warned >> there be dragons :) > > It seems platform_drm would cast disp->PlatformDisplay to a gbm_device > and use it if available; otherwise it always uses the first normal > node (/dev/dri/card0). > > Can it be assumed that if render nodes are available then it would > always be the first one (/dev/dri/renderD128)? Otherwise I still think > it is correct to run a for loop to try all the available render nodes > (renderD128..renderD191) > Ouch, so it seems that things are already in the "funny" lane. In that case I would like to withdraw my objection on the topic, considering that we already have a similar "not so elegant" solution.
Perhaps nvidia's GL extensions on the topic might provide a bit more flexibility ? Whereas for the "should we loop through all the rnodes vs open the first one", I don't have a strong preference for either one. Cheers, Emil P.S. For the future, please add a note as to which patch is the last one. Revision changelog/info would be appreciated as well. _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev