Bugfix: WaveActiveCountBits on glsl output. (#1488)

* Fix WaveActiveCountBits on glsl output.

* Fix warning `could not be inlined because the return instruction is not at the end of the function. This could be fixed by running merge-return before inlining.` from glslang - because we weren't including the CreateMergeReturnPasss on default optimization, and it's assumed in InlineExhaustivePass.

* Keep WaveActiveCountBits use the default WaveMask impl.

* Fix WaveCountBits calculation.
Use WaveActiveBallot instead of the _WaveActiveBallot.

1 files changed, 2 insertions, 0 deletions

diff --git a/source/slang-glslang/slang-glslang.cpp b/source/slang-glslang/slang-glslang.cpp
index 1c756cb6c..80087997c 100644
--- a/source/slang-glslang/slang-glslang.cpp
+++ b/source/slang-glslang/slang-glslang.cpp
@@ -168,6 +168,8 @@ static void glslang_optimizeSPIRV(std::vector<unsigned int>& spirv, spv_target_e
         break;
     case SLANG_OPTIMIZATION_LEVEL_DEFAULT:
         // Use a minimal set of performance settings
+        // If we run CreateInlineExhaustivePass, We need to run CreateMergeReturnPass first. 
+        optimizer.RegisterPass(spvtools::CreateMergeReturnPass());
         optimizer.RegisterPass(spvtools::CreateInlineExhaustivePass());
         optimizer.RegisterPass(spvtools::CreateAggressiveDCEPass());
         optimizer.RegisterPass(spvtools::CreatePrivateToLocalPass());