[Openmp-commits] [openmp] r273282 - Addition of debugger comments and whitespace

Jonathan Peyton via Openmp-commits openmp-commits at lists.llvm.org
Tue Jun 21 08:59:35 PDT 2016


Author: jlpeyton
Date: Tue Jun 21 10:59:34 2016
New Revision: 273282

URL: http://llvm.org/viewvc/llvm-project?rev=273282&view=rev
Log:
Addition of debugger comments and whitespace

The removal of legacy code to support long-deprecated debugger support library
resulted in some whitespace changes. Comments from that legacy code were made
public as they may be useful for other debuggers.

Patch by Olga Malysheva.

Differential Revision: http://reviews.llvm.org/D21391

Modified:
    openmp/trunk/runtime/src/dllexports
    openmp/trunk/runtime/src/kmp_runtime.c
    openmp/trunk/runtime/src/kmp_tasking.c

Modified: openmp/trunk/runtime/src/dllexports
URL: http://llvm.org/viewvc/llvm-project/openmp/trunk/runtime/src/dllexports?rev=273282&r1=273281&r2=273282&view=diff
==============================================================================
--- openmp/trunk/runtime/src/dllexports (original)
+++ openmp/trunk/runtime/src/dllexports Tue Jun 21 10:59:34 2016
@@ -186,15 +186,15 @@
         __kmp_omp_debug_struct_info             DATA
     %endif
 
-        # Symbols for MS mutual detection:
-        _You_must_link_with_exactly_one_OpenMP_library    DATA
-        _You_must_link_with_Intel_OpenMP_library          DATA
-        %ifdef msvc_compat
-            _You_must_link_with_Microsoft_OpenMP_library  DATA
-        %endif
+    # Symbols for MS mutual detection:
+    _You_must_link_with_exactly_one_OpenMP_library    DATA
+    _You_must_link_with_Intel_OpenMP_library          DATA
+    %ifdef msvc_compat
+        _You_must_link_with_Microsoft_OpenMP_library  DATA
+    %endif
 
-        __kmp_wait_64
-        __kmp_release_64
+    __kmp_wait_64
+    __kmp_release_64
 
 
 #    VT_getthid                              1

Modified: openmp/trunk/runtime/src/kmp_runtime.c
URL: http://llvm.org/viewvc/llvm-project/openmp/trunk/runtime/src/kmp_runtime.c?rev=273282&r1=273281&r2=273282&view=diff
==============================================================================
--- openmp/trunk/runtime/src/kmp_runtime.c (original)
+++ openmp/trunk/runtime/src/kmp_runtime.c Tue Jun 21 10:59:34 2016
@@ -531,7 +531,6 @@ static void __kmp_fini_allocator() {}
 #ifdef KMP_DYNAMIC_LIB
 # if KMP_OS_WINDOWS
 
-
 static void
 __kmp_reset_lock( kmp_bootstrap_lock_t* lck ) {
     // TODO: Change to __kmp_break_bootstrap_lock().

Modified: openmp/trunk/runtime/src/kmp_tasking.c
URL: http://llvm.org/viewvc/llvm-project/openmp/trunk/runtime/src/kmp_tasking.c?rev=273282&r1=273281&r2=273282&view=diff
==============================================================================
--- openmp/trunk/runtime/src/kmp_tasking.c (original)
+++ openmp/trunk/runtime/src/kmp_tasking.c Tue Jun 21 10:59:34 2016
@@ -1420,6 +1420,7 @@ __kmpc_omp_taskwait( ident_t *loc_ref, k
         }
 #endif
 
+        // Debugger: The taskwait is active. Store location and thread encountered the taskwait.
 #if USE_ITT_BUILD
         // Note: These values are used by ITT events as well.
 #endif /* USE_ITT_BUILD */
@@ -1452,6 +1453,7 @@ __kmpc_omp_taskwait( ident_t *loc_ref, k
 #endif /* USE_ITT_BUILD */
 
         // GEH TODO: shouldn't we have some sort of OMPRAP API calls here to mark end of wait?
+        // Debugger:  The taskwait is completed. Location remains, but thread is negated.
         taskdata->td_taskwait_thread = - taskdata->td_taskwait_thread;
 
 #if OMPT_SUPPORT && OMPT_TRACE
@@ -1494,6 +1496,7 @@ __kmpc_omp_taskyield( ident_t *loc_ref,
         thread = __kmp_threads[ gtid ];
         taskdata = thread -> th.th_current_task;
         // Should we model this as a task wait or not?
+        // Debugger: The taskwait is active. Store location and thread encountered the taskwait.
 #if USE_ITT_BUILD
         // Note: These values are used by ITT events as well.
 #endif /* USE_ITT_BUILD */
@@ -1521,6 +1524,7 @@ __kmpc_omp_taskyield( ident_t *loc_ref,
 #endif /* USE_ITT_BUILD */
 
         // GEH TODO: shouldn't we have some sort of OMPRAP API calls here to mark end of wait?
+        // Debugger:  The taskwait is completed. Location remains, but thread is negated.
         taskdata->td_taskwait_thread = - taskdata->td_taskwait_thread;
     }
 




More information about the Openmp-commits mailing list