On Sat, Jun 14, 2014 at 07:00:52AM +0100, Chris Wilson wrote: > This should enable accurate frontbuffer tracking and keep PSR disabled > whilst (and only while) the scanout is being directly accessed either by > the CPU or by the GPU. The important difference here is that the > tracking is more accurate and we check in the re-enable worker whether > the fb is still being written to. This should allow legacy display > servers to use PSR effectively, without breaking the special cases where > the scanout is being written to directly and never flushed (for example > after the display server is banned from the GPU). It also has the > side-effect of working correctly with fbcon, by disabling PSR whilst > fbcon is active. Nope. userpsace messes up fbcon tracking by knocking it out of the write domain by reading from it with the GPU. I think the correct answer here is just to fix fbcon to do the correct domain tracking, and then it a small step to adding the 100 lines or so to fully accelerate fbcon. -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx