On 01.05.2014 07:49, Alexandre Courbot wrote: > Agreed. Besides I hope the day won't come where we have to go through > 2 layers of memory translation for the GPU... That's actually the method of operation that gives us the best performance. GPU likes big pages, and without IOMMU translation you'd have a hard time finding enough contiguous physical memory. Terje _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/dri-devel