From: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx> This series mainly fixes the DDB allocation for cursor on skl+. Currently that is totally busted. In an effort to optimize cursor behaviour I'm also changing which plane watermarks we choose to disable. This was inspired by the 2x-cursor-vs-flip fails, but I don't see how this would help there as that test should not be changing the plane configuration in any significant way during the test (it should only move the cursor and perform page flips). And finally I sprinkled the usual amount of cleanups on top. Cc: Neel Desai <neel.desai@xxxxxxxxx> Ville Syrjälä (9) drm/i915: Accept alloc_size == blocks drm/i915: Don't pass plane state to skl_compute_plane_wm() drm/i915: Extract skl_compute_wm_params() drm/i915: Allocate enough DDB for the cursor drm/i915: Make sure cursor has enough ddb for the selected wm level drm/i915: Keep plane watermarks enabled more aggressively drm/i915: Move some variables to tighter scope drm/i915: Don't pass pipe_wm around so much drm/i915: Inline skl_build_pipe_wm() into its only caller drivers/gpu/drm/i915/intel_pm.c | 230 ++++++++++++++++++-------------- 1 file changed, 133 insertions(+), 97 deletions(-) -- 2.19.2 _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
