[clang] [llvm] [Offload][CUDA] Allow CUDA kernels to use LLVM/Offload (PR #94549)

[Jan Patrick Lehr via llvm-commits]

================
@@ -1199,7 +1244,9 @@ llvm::Function *CGNVCUDARuntime::finalizeModule() {
     }
     return nullptr;
   }
-  if (CGM.getLangOpts().OffloadingNewDriver && RelocatableDeviceCode)
+  if (CGM.getLangOpts().OffloadViaLLVM)
+    createOffloadingEntries();
+  else if (CGM.getLangOpts().OffloadingNewDriver && RelocatableDeviceCode)
----------------
jplehr wrote:

Is this calling the same target in both cases? Why is it distinguished?

https://github.com/llvm/llvm-project/pull/94549