===================================================================
@@ -204,6 +204,13 @@
when others => null;
end case;
+ -- Nothing to do for the identifier in an object renaming declaration,
+ -- the renaming itself does not need atomic syncrhonization.
+
+ if Nkind (Parent (N)) = N_Object_Renaming_Declaration then
+ return;
+ end if;
+
-- Go ahead and set the flag
Set_Atomic_Sync_Required (N);
===================================================================
@@ -1344,6 +1344,13 @@
Set_Is_Volatile (Id, Is_Volatile_Object (Nam));
+ -- Also copy settings of Is_Atomic and Is_Independent
+
+ if Is_Entity_Name (Nam) then
+ Set_Is_Atomic (Id, Is_Atomic (Entity (Nam)));
+ Set_Is_Independent (Id, Is_Independent (Entity (Nam)));
+ end if;
+
-- Treat as volatile if we just set the Volatile flag
if Is_Volatile (Id)
This update corrects two problems in the handling of Atomic. First we do not need Atomic_Synchronization for an object renaming declaration. Second, when we do have a renaming of an atomic object, the renaming object should be marked as atomic. Compiling this test: 1. package Renamed_Atomic is 2. 3. I : Integer; 4. pragma Atomic (I); 5. 6. function Get_I return Integer; 7. procedure Set_I (Val : Integer); 8. 9. J : Integer renames I; 10. 11. function Get_J return Integer; 12. procedure Set_J (Val : Integer); 13. 14. end Renamed_Atomic; We do not want Is_Atomic set on the identifier reference in line 9, but we do want Is_Atomic set on the entity for J. So if we use -gnatdt to generate a tree file, it should have two Is_Atomic references, one for the entity I and one for the entity J. Tested on x86_64-pc-linux-gnu, committed on trunk 2015-05-21 Robert Dewar <dewar@adacore.com> * exp_util.adb (Activate_Atomic_Synchronization): Do not set Atomic_Sync_Required for an object renaming declaration. * sem_ch8.adb (Analyze_Object_Renaming): Copy Is_Atomic and Is_Independent to renaming object.