[flang][OpenMP][NFC] test the current private dealloc runtime calls #124017
Conversation
It looks like in most cases we still don't make calls to deallocate allocatable members of derived types which have been privatized. This is just intended to add a test for the one case where we do, to make sure this doesn't regress with my upcoming changes.
llvmbot
added
flang
|
@llvm/pr-subscribers-flang-fir-hlfir Author: Tom Eccles (tblah) ChangesIt looks like in most cases we still don't make calls to deallocate allocatable members of derived types which have been privatized. This is just intended to add a test for the one case where we do, to make sure this doesn't regress with my upcoming changes. Full diff: https://github.com/llvm/llvm-project/pull/124017.diff 1 Files Affected:
diff --git a/flang/test/Lower/OpenMP/derived-type-allocatable.f90 b/flang/test/Lower/OpenMP/derived-type-allocatable.f90
index 2dc4e20f27af21..1d6e22212eedd0 100644
--- a/flang/test/Lower/OpenMP/derived-type-allocatable.f90
+++ b/flang/test/Lower/OpenMP/derived-type-allocatable.f90
@@ -24,6 +24,9 @@ module m1
!CHECK-LABEL: omp.private {type = private} @_QMm1Ftest_array_of_allocs
!CHECK: fir.call @_FortranAInitializeClone
!CHECK-NEXT: omp.yield
+!CHECK: } dealloc {
+!CHECK: fir.call @_FortranAAllocatableDeallocate
+!CHECK: omp.yield
!CHECK-LABEL: omp.private {type = firstprivate} @_QMm1Ftest_array
!CHECK-NOT: fir.call @_FortranAInitializeClone
|
|
@llvm/pr-subscribers-flang-openmp Author: Tom Eccles (tblah) ChangesIt looks like in most cases we still don't make calls to deallocate allocatable members of derived types which have been privatized. This is just intended to add a test for the one case where we do, to make sure this doesn't regress with my upcoming changes. Full diff: https://github.com/llvm/llvm-project/pull/124017.diff 1 Files Affected:
diff --git a/flang/test/Lower/OpenMP/derived-type-allocatable.f90 b/flang/test/Lower/OpenMP/derived-type-allocatable.f90
index 2dc4e20f27af21..1d6e22212eedd0 100644
--- a/flang/test/Lower/OpenMP/derived-type-allocatable.f90
+++ b/flang/test/Lower/OpenMP/derived-type-allocatable.f90
@@ -24,6 +24,9 @@ module m1
!CHECK-LABEL: omp.private {type = private} @_QMm1Ftest_array_of_allocs
!CHECK: fir.call @_FortranAInitializeClone
!CHECK-NEXT: omp.yield
+!CHECK: } dealloc {
+!CHECK: fir.call @_FortranAAllocatableDeallocate
+!CHECK: omp.yield
!CHECK-LABEL: omp.private {type = firstprivate} @_QMm1Ftest_array
!CHECK-NOT: fir.call @_FortranAInitializeClone
|
It looks like in most cases we still don't make calls to deallocate allocatable members of derived types which have been privatized.
This is just intended to add a test for the one case where we do, to make sure this doesn't regress with my upcoming changes.