Index: openmp/runtime/test/ompt/synchronization/ordered_dependences.c
===================================================================
--- openmp/runtime/test/ompt/synchronization/ordered_dependences.c
+++ openmp/runtime/test/ompt/synchronization/ordered_dependences.c
@@ -25,39 +25,37 @@
 // CHECK: 0: NULL_POINTER=[[NULL:.*$]]
 
 // CHECK: {{^}}[[MASTER:[0-9]+]]: ompt_event_loop_begin:
-// CHECK-SAME: parallel_id={{[0-9]+}}, parent_task_id=[[IMPL_TASK:[0-9]+]],
+// CHECK-SAME: parallel_id={{[0-9]+}}, parent_task_id=[[ITASK:[0-9]+]],
 
-// CHECK: {{^}}[[MASTER]]: ompt_event_dependences: task_id=[[IMPL_TASK]],
+// CHECK: {{^}}[[MASTER]]: ompt_event_dependences: task_id=[[ITASK]],
 // CHECK-SAME: deps=[(0, ompt_dependence_type_source), (0,
 // CHECK-SAME: ompt_dependence_type_source)], ndeps=2
 
-// CHECK: {{^}}[[MASTER]]: ompt_event_dependences: task_id=[[IMPL_TASK]],
+// CHECK: {{^}}[[MASTER]]: ompt_event_dependences: task_id=[[ITASK]],
 // CHECK-SAME: deps=[(0, ompt_dependence_type_sink), (0,
 // CHECK-SAME: ompt_dependence_type_sink)], ndeps=2
 
-// CHECK: {{^}}[[MASTER]]: ompt_event_dependences: task_id=[[IMPL_TASK]],
+// CHECK: {{^}}[[MASTER]]: ompt_event_dependences: task_id=[[ITASK]],
 // CHECK-SAME: deps=[(0, ompt_dependence_type_source), (1,
 // CHECK-SAME: ompt_dependence_type_source)], ndeps=2
 
 // CHECK: {{^}}[[WORKER:[0-9]+]]: ompt_event_loop_begin:
-// CHECK-SAME: parallel_id={{[0-9]+}}, parent_task_id=[[IMPL_TASK:[0-9]+]],
+// CHECK-SAME: parallel_id={{[0-9]+}}, parent_task_id=[[ITASK:[0-9]+]],
 
-// CHECK: {{^}}[[WORKER]]: ompt_event_dependences: task_id=[[IMPL_TASK]],
+// CHECK: {{^}}[[WORKER]]: ompt_event_dependences: task_id=[[ITASK]],
 // CHECK-SAME: deps=[(0, ompt_dependence_type_sink), (0,
 // CHECK-SAME: ompt_dependence_type_sink)], ndeps=2
 
-// CHECK: {{^}}[[WORKER]]: ompt_event_dependences: task_id=[[IMPL_TASK]],
+// CHECK: {{^}}[[WORKER]]: ompt_event_dependences: task_id=[[ITASK]],
 // CHECK-SAME: deps=[(1, ompt_dependence_type_source), (0,
 // CHECK-SAME: ompt_dependence_type_source)], ndeps=2
 
-// CHECK: {{^}}[[WORKER]]: ompt_event_dependences: task_id=[[IMPL_TASK]],
-// CHECK-SAME: deps=[(0, ompt_dependence_type_sink), (1,
-// CHECK-SAME: ompt_dependence_type_sink)], ndeps=2
+// either can be first for last iteration
 
-// CHECK: {{^}}[[WORKER]]: ompt_event_dependences: task_id=[[IMPL_TASK]],
-// CHECK-SAME: deps=[(1, ompt_dependence_type_sink), (0,
-// CHECK-SAME: ompt_dependence_type_sink)], ndeps=2
+// CHECK-DAG: [[ITASK]]{{.*}}deps=[(0{{.*}}sink), (1,{{.*}}sink)]
+
+// CHECK-DAG: [[ITASK]]{{.*}}deps=[(1{{.*}}sink), (0,{{.*}}sink)]
 
-// CHECK: {{^}}[[WORKER]]: ompt_event_dependences: task_id=[[IMPL_TASK]],
+// CHECK: {{^}}[[WORKER]]: ompt_event_dependences: task_id=[[ITASK]],
 // CHECK-SAME: deps=[(1, ompt_dependence_type_source), (1,
 // CHECK-SAME: ompt_dependence_type_source)], ndeps=2
Index: openmp/runtime/test/ompt/tasks/dependences.c
===================================================================
--- openmp/runtime/test/ompt/tasks/dependences.c
+++ openmp/runtime/test/ompt/tasks/dependences.c
@@ -24,14 +24,6 @@
       print_fuzzy_address(1);
       print_ids(0);
 
-#pragma omp task depend(mutexinoutset : x)
-      {
-        x++;
-        delay(100);
-      }
-      print_fuzzy_address(2);
-      print_ids(0);
-
 #pragma omp task depend(in : x)
       { x = -1; }
       print_ids(0);
@@ -86,28 +78,10 @@
 
 // CHECK: {{^}}[[MASTER_ID]]: ompt_event_dependences:
 // CHECK-SAME: task_id=[[SECOND_TASK]], deps=[([[ADDRX]],
-// CHECK-SAME: ompt_dependence_type_mutexinoutset)], ndeps=1
-
-// CHECK: {{^}}[[MASTER_ID]]: ompt_event_task_dependence_pair:
-// CHECK-SAME: first_task_id=[[FIRST_TASK]], second_task_id=[[SECOND_TASK]]
-
-// CHECK: {{^}}[[MASTER_ID]]: fuzzy_address={{.*}}[[RETURN_ADDRESS]]
-// CHECK: {{^}}[[MASTER_ID]]: task level 0: parallel_id=[[PARALLEL_ID]],
-// CHECK-SAME: task_id=[[IMPLICIT_TASK_ID]], exit_frame=[[EXIT]],
-// CHECK-SAME: reenter_frame=[[NULL]]
-
-// CHECK: {{^}}[[MASTER_ID]]: ompt_event_task_create:
-// CHECK-SAME: parent_task_id={{[0-9]+}}, parent_task_frame.exit=[[EXIT]],
-// CHECK-SAME: parent_task_frame.reenter={{0x[0-f]+}},
-// CHECK-SAME: new_task_id=[[THIRD_TASK:[0-f]+]], codeptr_ra={{0x[0-f]+}},
-// CHECK-SAME: task_type=ompt_task_explicit=4, has_dependences=yes
-
-// CHECK: {{^}}[[MASTER_ID]]: ompt_event_dependences:
-// CHECK-SAME: task_id=[[THIRD_TASK]], deps=[([[ADDRX]],
 // CHECK-SAME: ompt_dependence_type_in)], ndeps=1
 
 // CHECK: {{^}}[[MASTER_ID]]: ompt_event_task_dependence_pair:
-// CHECK-SAME: first_task_id=[[SECOND_TASK]], second_task_id=[[THIRD_TASK]]
+// CHECK-SAME: first_task_id=[[FIRST_TASK]], second_task_id=[[SECOND_TASK]]
 
 // CHECK: {{^}}[[MASTER_ID]]: task level 0: parallel_id=[[PARALLEL_ID]],
 // CHECK-SAME: task_id=[[IMPLICIT_TASK_ID]], exit_frame=[[EXIT]],
Index: openmp/runtime/test/ompt/tasks/dependences_mtxinoutset.c
===================================================================
--- /dev/null
+++ openmp/runtime/test/ompt/tasks/dependences_mtxinoutset.c
@@ -0,0 +1,118 @@
+// RUN: %libomp-compile-and-run | %sort-threads | FileCheck %s
+// REQUIRES: ompt
+// UNSUPPORTED: gcc, clang-4, clang-5, clang-6, clang-7, clang-8
+
+// GCC does not pass in mutexinoutset
+// clang 9 introduced codegen for mutexinoutset
+
+#include "callback.h"
+#include <omp.h>
+#include <math.h>
+#include <unistd.h>
+
+int main() {
+  int x = 0;
+#pragma omp parallel num_threads(2)
+  {
+#pragma omp master
+    {
+      print_ids(0);
+      printf("%" PRIu64 ": address of x: %p\n", ompt_get_thread_data()->value,
+             &x);
+#pragma omp task depend(out : x)
+      {
+        x++;
+        delay(100);
+      }
+      print_fuzzy_address(1);
+      print_ids(0);
+
+#pragma omp task depend(mutexinoutset : x)
+      {
+        x++;
+        delay(100);
+      }
+      print_fuzzy_address(2);
+      print_ids(0);
+
+#pragma omp task depend(in : x)
+      { x = -1; }
+      print_ids(0);
+    }
+  }
+
+  x++;
+
+  return 0;
+}
+
+// Check if libomp supports the callbacks for this test.
+// CHECK-NOT: {{^}}0: Could not register callback 'ompt_callback_task_create'
+// CHECK-NOT: {{^}}0: Could not register callback 'ompt_callback_dependences'
+// CHECK-NOT: {{^}}0: Could not register callback 'ompt_callback_task_depende
+
+// CHECK: {{^}}0: NULL_POINTER=[[NULL:.*$]]
+
+// make sure initial data pointers are null
+// CHECK-NOT: 0: new_task_data initially not null
+
+// CHECK: {{^}}[[MASTER_ID:[0-9]+]]: ompt_event_implicit_task_begin:
+// CHECK-SAME: parallel_id=[[PARALLEL_ID:[0-9]+]],
+// CHECK-SAME: task_id=[[IMPLICIT_TASK_ID:[0-9]+]]
+
+// CHECK: {{^}}[[MASTER_ID]]: task level 0: parallel_id=[[PARALLEL_ID]],
+// CHECK-SAME: task_id=[[IMPLICIT_TASK_ID]], exit_frame=[[EXIT:0x[0-f]+]],
+// CHECK-SAME: reenter_frame=[[NULL]]
+
+// CHECK: {{^}}[[MASTER_ID]]: address of x: [[ADDRX:0x[0-f]+]]
+// CHECK: {{^}}[[MASTER_ID]]: ompt_event_task_create:
+// CHECK-SAME: parent_task_id={{[0-9]+}}, parent_task_frame.exit=[[EXIT]],
+// CHECK-SAME: parent_task_frame.reenter={{0x[0-f]+}},
+// CHECK-SAME: new_task_id=[[FIRST_TASK:[0-f]+]],
+// CHECK-SAME: codeptr_ra=[[RETURN_ADDRESS:0x[0-f]+]]{{[0-f][0-f]}},
+// CHECK-SAME: task_type=ompt_task_explicit=4, has_dependences=yes
+
+// CHECK: {{^}}[[MASTER_ID]]: ompt_event_dependences:
+// CHECK-SAME: task_id=[[FIRST_TASK]], deps=[([[ADDRX]],
+// CHECK-SAME: ompt_dependence_type_inout)], ndeps=1
+
+// CHECK: {{^}}[[MASTER_ID]]: fuzzy_address={{.*}}[[RETURN_ADDRESS]]
+// CHECK: {{^}}[[MASTER_ID]]: task level 0: parallel_id=[[PARALLEL_ID]],
+// CHECK-SAME: task_id=[[IMPLICIT_TASK_ID]], exit_frame=[[EXIT]],
+// CHECK-SAME: reenter_frame=[[NULL]]
+
+// CHECK: {{^}}[[MASTER_ID]]: ompt_event_task_create:
+// CHECK-SAME: parent_task_id={{[0-9]+}}, parent_task_frame.exit=[[EXIT]],
+// CHECK-SAME: parent_task_frame.reenter={{0x[0-f]+}},
+// CHECK-SAME: new_task_id=[[SECOND_TASK:[0-f]+]], codeptr_ra={{0x[0-f]+}},
+// CHECK-SAME: codeptr_ra=[[RETURN_ADDRESS:0x[0-f]+]]{{[0-f][0-f]}},
+// CHECK-SAME: task_type=ompt_task_explicit=4, has_dependences=yes
+
+// CHECK: {{^}}[[MASTER_ID]]: ompt_event_dependences:
+// CHECK-SAME: task_id=[[SECOND_TASK]], deps=[([[ADDRX]],
+// CHECK-SAME: ompt_dependence_type_mutexinoutset)], ndeps=1
+
+// CHECK: {{^}}[[MASTER_ID]]: ompt_event_task_dependence_pair:
+// CHECK-SAME: first_task_id=[[FIRST_TASK]], second_task_id=[[SECOND_TASK]]
+
+// CHECK: {{^}}[[MASTER_ID]]: fuzzy_address={{.*}}[[RETURN_ADDRESS]]
+// CHECK: {{^}}[[MASTER_ID]]: task level 0: parallel_id=[[PARALLEL_ID]],
+// CHECK-SAME: task_id=[[IMPLICIT_TASK_ID]], exit_frame=[[EXIT]],
+// CHECK-SAME: reenter_frame=[[NULL]]
+
+// CHECK: {{^}}[[MASTER_ID]]: ompt_event_task_create:
+// CHECK-SAME: parent_task_id={{[0-9]+}}, parent_task_frame.exit=[[EXIT]],
+// CHECK-SAME: parent_task_frame.reenter={{0x[0-f]+}},
+// CHECK-SAME: new_task_id=[[THIRD_TASK:[0-f]+]], codeptr_ra={{0x[0-f]+}},
+// CHECK-SAME: task_type=ompt_task_explicit=4, has_dependences=yes
+
+// CHECK: {{^}}[[MASTER_ID]]: ompt_event_dependences:
+// CHECK-SAME: task_id=[[THIRD_TASK]], deps=[([[ADDRX]],
+// CHECK-SAME: ompt_dependence_type_in)], ndeps=1
+
+// CHECK: {{^}}[[MASTER_ID]]: ompt_event_task_dependence_pair:
+// CHECK-SAME: first_task_id=[[SECOND_TASK]], second_task_id=[[THIRD_TASK]]
+
+// CHECK: {{^}}[[MASTER_ID]]: task level 0: parallel_id=[[PARALLEL_ID]],
+// CHECK-SAME: task_id=[[IMPLICIT_TASK_ID]], exit_frame=[[EXIT]],
+// CHECK-SAME: reenter_frame=[[NULL]]
Index: openmp/runtime/test/ompt/tasks/task_early_fulfill.c
===================================================================
--- openmp/runtime/test/ompt/tasks/task_early_fulfill.c
+++ openmp/runtime/test/ompt/tasks/task_early_fulfill.c
@@ -1,8 +1,8 @@
 // RUN: %libomp-compile -fopenmp-version=50 && env OMP_NUM_THREADS='3' \
 // RUN:    %libomp-run | %sort-threads | FileCheck %s
 
-// Checked gcc 9.2 still does not support detach clause on task construct.
-// UNSUPPORTED: gcc-4, gcc-5, gcc-6, gcc-7, gcc-8, gcc-9
+// Checked gcc 10.1 still does not support detach clause on task construct.
+// UNSUPPORTED: gcc-4, gcc-5, gcc-6, gcc-7, gcc-8, gcc-9, gcc-10
 // clang supports detach clause since version 11.
 // UNSUPPORTED: clang-10, clang-9, clang-8, clang-7
 // icc compiler does not support detach clause.
Index: openmp/runtime/test/ompt/tasks/task_late_fulfill.c
===================================================================
--- openmp/runtime/test/ompt/tasks/task_late_fulfill.c
+++ openmp/runtime/test/ompt/tasks/task_late_fulfill.c
@@ -1,8 +1,8 @@
 // RUN: %libomp-compile -fopenmp-version=50 && env OMP_NUM_THREADS='3' \
 // RUN:    %libomp-run | %sort-threads | FileCheck %s
 
-// Checked gcc 9.2 still does not support detach clause on task construct.
-// UNSUPPORTED: gcc-4, gcc-5, gcc-6, gcc-7, gcc-8, gcc-9
+// Checked gcc 10.1 still does not support detach clause on task construct.
+// UNSUPPORTED: gcc-4, gcc-5, gcc-6, gcc-7, gcc-8, gcc-9, gcc-10
 // clang supports detach clause since version 11.
 // UNSUPPORTED: clang-10, clang-9, clang-8, clang-7
 // icc compiler does not support detach clause.
Index: openmp/runtime/test/ompt/worksharing/for/runtime.c
===================================================================
--- openmp/runtime/test/ompt/worksharing/for/runtime.c
+++ openmp/runtime/test/ompt/worksharing/for/runtime.c
@@ -1,5 +1,7 @@
 // RUN: %libomp-compile-and-run | %sort-threads | FileCheck %S/base.h
 // REQUIRES: ompt
+// gcc 9/10 use GOMP_parallel_loop_maybe_nonmonotonic_runtime, not implemented
+// XFAIL: gcc-9, gcc-10
 
 #define SCHEDULE runtime
 #include "base.h"
Index: openmp/runtime/test/ompt/worksharing/for/runtime_serialized.c
===================================================================
--- openmp/runtime/test/ompt/worksharing/for/runtime_serialized.c
+++ openmp/runtime/test/ompt/worksharing/for/runtime_serialized.c
@@ -1,5 +1,7 @@
 // RUN: %libomp-compile-and-run | %sort-threads | FileCheck %S/base_serialized.h
 // REQUIRES: ompt
+// gcc 9/10 use GOMP_parallel_loop_maybe_nonmonotonic_runtime, not implemented
+// XFAIL: gcc-9, gcc-10
 
 #define SCHEDULE runtime
 #include "base_serialized.h"
Index: openmp/runtime/test/ompt/worksharing/for/runtime_split.c
===================================================================
--- openmp/runtime/test/ompt/worksharing/for/runtime_split.c
+++ openmp/runtime/test/ompt/worksharing/for/runtime_split.c
@@ -3,5 +3,8 @@
 // REQUIRES: ompt
 // UNSUPPORTED: gcc-4, gcc-5, gcc-6, gcc-7
 
+// gcc 9/10 use GOMP_parallel_loop_maybe_nonmonotonic_runtime, not implemented
+// XFAIL: gcc-9, gcc-10
+
 #define SCHEDULE runtime
 #include "base_split.h"