[clang] [CIR][CUDA] Support built-in CUDA texture type (PR #214781)

[Andy Kaylor via cfe-commits]

================
@@ -3861,6 +3694,9 @@ static void prepareTypeConverter(mlir::LLVMTypeConverter &converter,
     assert(!cir::MissingFeatures::addressSpace());
     return mlir::LLVM::LLVMPointerType::get(type.getContext());
   });
+  converter.addConversion([&](cir::CUDADeviceTextureType type) -> mlir::Type {
----------------
andykaylor wrote:

Yes, a follow-up PR for the surface type would be good.

https://github.com/llvm/llvm-project/pull/214781