diff options
| author | Samuel Pitoiset <[email protected]> | 2019-09-06 10:34:35 +0200 |
|---|---|---|
| committer | Samuel Pitoiset <[email protected]> | 2019-09-06 17:25:28 +0200 |
| commit | c6be5cefba7e9480f103c2ce2794a67ed9fa33eb (patch) | |
| tree | 250578cbecd92ef47057f1cc2060cd660472cf21 | |
| parent | b1a872f0c0f03abfeb63773b509d633bb4879d24 (diff) | |
radv/gfx10: make use the output usage mask when exporting NGG GS params
It shouldn't matter much because output varyings should have been
compacted during NIR shader linking but it mirrors what the driver
does when emitting NGG GS vertex parameters.
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
| -rw-r--r-- | src/amd/vulkan/radv_nir_to_llvm.c | 11 |
1 files changed, 8 insertions, 3 deletions
diff --git a/src/amd/vulkan/radv_nir_to_llvm.c b/src/amd/vulkan/radv_nir_to_llvm.c index 10cc82eb8d9..0c13860fe28 100644 --- a/src/amd/vulkan/radv_nir_to_llvm.c +++ b/src/amd/vulkan/radv_nir_to_llvm.c @@ -3489,16 +3489,21 @@ static void gfx10_ngg_gs_emit_epilogue_2(struct radv_shader_context *ctx) unsigned out_idx = 0; gep_idx[1] = ctx->ac.i32_0; for (unsigned i = 0; i < AC_LLVM_MAX_OUTPUTS; ++i) { + unsigned output_usage_mask = + ctx->shader_info->gs.output_usage_mask[i]; + int length = util_last_bit(output_usage_mask); + if (!(ctx->output_mask & (1ull << i))) continue; outputs[noutput].slot_name = i; outputs[noutput].slot_index = i == VARYING_SLOT_CLIP_DIST1; - - outputs[noutput].usage_mask = ctx->shader_info->gs.output_usage_mask[i]; - int length = util_last_bit(outputs[noutput].usage_mask); + outputs[noutput].usage_mask = output_usage_mask; for (unsigned j = 0; j < length; j++, out_idx++) { + if (!(output_usage_mask & (1 << j))) + continue; + gep_idx[2] = LLVMConstInt(ctx->ac.i32, out_idx, false); tmp = LLVMBuildGEP(builder, vertexptr, gep_idx, 3, ""); tmp = LLVMBuildLoad(builder, tmp, ""); |