[clang] [SYCL] The sycl_kernel_entry_point attribute. (PR #111389)

[Erich Keane via cfe-commits]

================
@@ -14296,6 +14296,31 @@ void ASTContext::getFunctionFeatureMap(llvm::StringMap<bool> &FeatureMap,
   }
 }
 
+static SYCLKernelInfo BuildSYCLKernelInfo(ASTContext &Context,
+                                          CanQualType KernelNameType,
+                                          const FunctionDecl *FD) {
+  return { KernelNameType, FD };
+}
+
+void ASTContext::registerSYCLEntryPointFunction(FunctionDecl *FD) {
+  assert(!FD->isInvalidDecl());
+  assert(!FD->isDependentContext());
+
+  const auto *SKEPAttr = FD->getAttr<SYCLKernelEntryPointAttr>();
+  assert(SKEPAttr && "Missing sycl_kernel_entry_point attribute");
+
+  CanQualType KernelNameType = getCanonicalType(SKEPAttr->getKernelName());
+  auto IT = SYCLKernels.find(KernelNameType);
+  if (IT != SYCLKernels.end()) {
+    if (!declaresSameEntity(FD, IT->second.GetKernelEntryPointDecl()))
+      llvm::report_fatal_error("SYCL kernel name conflict");
----------------
erichkeane wrote:

I'm really not a fan of 'temporary' fatal errors, but I can withhold judgement until later I guess.  

That said, why is THIS function where the diagnostic should happen?  Should this be an assert instead? It would be weird for `ASTContext` to diagnose. I would expect the diagnostic to happen earlier, so an 'assert' seems more reasonable here.  Also, it would allow us to then only check the `declaresSameEntity` (a reasonably expensive function) fewer times.

https://github.com/llvm/llvm-project/pull/111389