In addition to first version there is now also path for miptree updownsampling. While the relevant piglit tests don't crash on BDW anymore, they do not pass either - it seems that there are still SRGB blits needed that on previous hardware are covered by blorp. Using the stencil texturing support on IVB I was able to check that the new stencil updownsampling logic does not regress these tests.
Topi Pohjolainen (8): i965/wm: Surface state overrides for configuring w-tiled as y-tiled i965/gen8: Surface state overriding for stencil i965: Extend brw_get_rb_for_first_slice() for specified level/layer meta: Refactor state save/restore for framebuffer texture blits i965/meta: Stencil blits i965/fb: Use meta path for stencil blits i965/meta: Stencil blit for miptree updownsampling i965/fb: Use meta path for stencil up/downsampling src/mesa/drivers/common/meta.h | 15 + src/mesa/drivers/common/meta_blit.c | 59 ++- src/mesa/drivers/dri/i965/Makefile.sources | 1 + src/mesa/drivers/dri/i965/brw_context.h | 14 + src/mesa/drivers/dri/i965/brw_meta_stencil_blit.c | 526 ++++++++++++++++++++++ src/mesa/drivers/dri/i965/brw_meta_updownsample.c | 32 +- src/mesa/drivers/dri/i965/brw_state.h | 6 + src/mesa/drivers/dri/i965/brw_wm_surface_state.c | 24 + src/mesa/drivers/dri/i965/gen8_surface_state.c | 34 +- src/mesa/drivers/dri/i965/intel_fbo.c | 9 + src/mesa/drivers/dri/i965/intel_mipmap_tree.c | 9 +- 11 files changed, 686 insertions(+), 43 deletions(-) create mode 100644 src/mesa/drivers/dri/i965/brw_meta_stencil_blit.c -- 1.8.3.1 _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
