From the Sky Lake PRM: "For SURFTYPE_CUBE: For Sampling Engine Surfaces and Typed Data Port Surfaces, the range of this field is [0,340], indicating the number of cube array elements (equal to the number of underlying 2D array elements divided by 6). For other surfaces, this field must be zero."
In other words, the depth field for cube maps is in number of cubes not number of 2-D slices so we need to divide by 6. It appears as if we've been doing this wrong ever since we first added cube map arrays for Sandy Bridge. We've also had a shader hack to divide the size Z dimension of cube maps by 6 in the textureSize call. This is completely bogus and the only reason for it is that we've been setting the depth six times too large. This little series fixes this. In order to keep things back-portable, patch 3 comes in two versions. Version (a) is based on pre-ISL and should be backportable to 12.0 or maybe even 11.2 or 11.1. Version (b) on the other hand is based on top of the ISL work and can be applied on master. Cc: Emil Velikov <emil.veli...@collabora.com> Jason Ekstrand (3): i965: Use intel_get_image_dims in alloc_texture_storage i965/miptree: Set logical_depth0 == 6 for cube maps i965: Correctly set RENDER_SURFACE_STATE::Depth for cube map textures src/mesa/drivers/dri/i965/brw_fs_nir.cpp | 21 +++++---------------- src/mesa/drivers/dri/i965/brw_wm_surface_state.c | 6 +++++- src/mesa/drivers/dri/i965/gen7_wm_surface_state.c | 3 ++- src/mesa/drivers/dri/i965/gen8_surface_state.c | 3 ++- src/mesa/drivers/dri/i965/intel_mipmap_tree.c | 14 ++++++++++---- src/mesa/drivers/dri/i965/intel_tex.c | 2 ++ 6 files changed, 26 insertions(+), 23 deletions(-) -- 2.5.0.400.gff86faf _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev