On Sun, Aug 10, 2014 at 07:55:46AM +0100, Chris Wilson wrote: > On Sat, Aug 09, 2014 at 12:29:17PM -0700, Ben Widawsky wrote: > > I didn't read the patch closely, but the problem is clear. It seems like > > the one hunk: > > > - if (use_cpu_reloc(obj)) > > > + if (use_cpu_reloc(obj) || !i915_gem_obj_ggtt_bound(obj)) > > > > would be sufficient to fix the problem. > > Right, I wanted this patch as later patches build upon the flags and > using the flags makes the code cleaner. This tweak was something that > occurred to me when reviewing the patch and looking at recent bug > reports. As it happens, that is the key to the patch. Setting obj->map_and_fenceable correctly makes that tweak redundant. -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx