It doesn't matter what we use as the MCS fetch will ignore it. Using
zero means we'll naver trigger a header because of it. We have an
optimization in the indirect message case for when the texture and
sampler are the same but we also have one for when the texture is
indirect and the sampler is an immediate that's just as efficient so
this generates no more code than we had before.
---
src/intel/compiler/brw_fs_visitor.cpp | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/intel/compiler/brw_fs_visitor.cpp
b/src/intel/compiler/brw_fs_visitor.cpp
index 51a0ca2374a..259a56245ca 100644
--- a/src/intel/compiler/brw_fs_visitor.cpp
+++ b/src/intel/compiler/brw_fs_visitor.cpp
@@ -42,7 +42,8 @@ fs_visitor::emit_mcs_fetch(const fs_reg &coordinate, unsigned
components,
fs_reg srcs[TEX_LOGICAL_NUM_SRCS];
srcs[TEX_LOGICAL_SRC_COORDINATE] = coordinate;
srcs[TEX_LOGICAL_SRC_SURFACE] = texture;
- srcs[TEX_LOGICAL_SRC_SAMPLER] = texture;
+ /* sampler is ignored for a MCS fetch so it doesn't matter */
+ srcs[TEX_LOGICAL_SRC_SAMPLER] = brw_imm_ud(0);
srcs[TEX_LOGICAL_SRC_COORD_COMPONENTS] = brw_imm_d(components);
srcs[TEX_LOGICAL_SRC_GRAD_COMPONENTS] = brw_imm_d(0);
--
2.20.1
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
