[Mesa-dev] [PATCH] radv/ac/llvm: trim texture return values
Dave Airlie
airlied at gmail.com
Wed Oct 26 01:04:00 UTC 2016
From: Dave Airlie <airlied at redhat.com>
The intrinsic engine asserts in llvm due to this,
as we put a vec4 into a vec1, and the next instruction
isn't expecting it.
So trim the vector at the end before inserting it.
Reported-by: Christoph Haag <haagch+mesadev at frickel.club>
Cc: "13.0" <mesa-stable at lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied at redhat.com>
---
src/amd/common/ac_nir_to_llvm.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index c3f3fe8..e4efcf6 100644
--- a/src/amd/common/ac_nir_to_llvm.c
+++ b/src/amd/common/ac_nir_to_llvm.c
@@ -3525,7 +3525,8 @@ static void visit_tex(struct nir_to_llvm_context *ctx, nir_tex_instr *instr)
LLVMValueRef z = LLVMBuildExtractElement(ctx->builder, result, two, "");
z = LLVMBuildSDiv(ctx->builder, z, six, "");
result = LLVMBuildInsertElement(ctx->builder, result, z, two, "");
- }
+ } else if (instr->dest.ssa.num_components != 4)
+ result = trim_vector(ctx, result, instr->dest.ssa.num_components);
write_result:
if (result) {
--
2.5.5
More information about the mesa-dev
mailing list