On Fri, Feb 13, 2015 at 10:59:42AM +0200, Ville Syrjälä wrote: > Hmm, except we don't do it for clear_range(), but I guess that's not a > huge issue, just means someone could clobber some other memory besides > the scratch page if accidentally writing to a cleared area of the ggtt. It's a very small hole, only since it is the ggtt the only user is the kernel. (Except for i915.enable_ppgtt=0.) It is possible to forgo that clear entirely (looks at the PD implementation for a current violator, and that is a good example for where the extra WC writes can be have easurable impact on synmark). -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
