================
@@ -59,6 +59,9 @@ enum class LangAS : unsigned {
// HLSL specific address spaces.
hlsl_groupshared,
+ // Vulkan specific address spaces.
+ vulkan_private,
+
----------------
s-perron wrote:
Do we need a new address space? This is not an issue specific to these input
and output variable. We need to figure out how to handle general static
variables. There are some tests that already consider this, but they are DXIL
only.
https://github.com/llvm/llvm-project/blob/94eebf721a2f8630412730f51d5071816a686ea0/clang/test/CodeGenHLSL/GlobalDestructors.hlsl#L45
We might need to figure that out first, and than have this PR build on top of
that. For DXIL, the static variables are marked as "internal global", and does
not use a different address space. Can we do the same? Is the DXIL codegen
their real solution or just a temporary implementation?
https://github.com/llvm/llvm-project/pull/116393
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
