[clang] Thread Safety Analysis: Support warning on taking address of guarded variables (PR #123063)

[Aaron Puchert via cfe-commits]

================
@@ -1983,11 +1983,21 @@ class ThreadSafetyReporter : public clang::threadSafety::ThreadSafetyHandler {
 
   void handleNoMutexHeld(const NamedDecl *D, ProtectedOperationKind POK,
                          AccessKind AK, SourceLocation Loc) override {
-    assert((POK == POK_VarAccess || POK == POK_VarDereference) &&
-           "Only works for variables");
-    unsigned DiagID = POK == POK_VarAccess?
-                        diag::warn_variable_requires_any_lock:
-                        diag::warn_var_deref_requires_any_lock;
+    unsigned DiagID = 0;
+    switch (POK) {
+    case POK_VarAccess:
+      DiagID = diag::warn_variable_requires_any_lock;
+      break;
+    case POK_VarDereference:
+      DiagID = diag::warn_var_deref_requires_any_lock;
+      break;
+    case POK_AddressOf:
+      DiagID = diag::warn_addressof_requires_any_lock;
+      break;
+    default:
+      assert(false && "Only works for variables");
+      break;
----------------
aaronpuchert wrote:

Wouldn't `llvm_unreachable` be the standard idiom here?

https://github.com/llvm/llvm-project/pull/123063