[Flang][OpenMP][MLIR] Check for presence of Box type before emitting store in MapInfoFinalization pass #135477
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…store in MapInfoFinalization pass Currently we don't check for the presence of descriptor/BoxTypes before emitting stores which lower to memcpys, the issue with this is that users can have optional arguments, where they don't provide an input, making the argument effectively null. This can still be mapped and this causes issues at the moment as we'll emit a memcpy for function arguments to store to a local variable for certain edge cases, when we perform this memcpy on a null input, we cause a segfault at runtime. The fix to this is to simply create a branch around the store that checks if the data we're copying from is actually present. If it is, we procede with the store, if it isn't we skip it.
llvmbot
added
flang
Member
|
@llvm/pr-subscribers-flang-openmp @llvm/pr-subscribers-flang-fir-hlfir Author: None (agozillon) ChangesCurrently we don't check for the presence of descriptor/BoxTypes before emitting stores which lower to memcpys, the issue with this is that users can have optional arguments, where they don't provide an input, making the argument effectively null. This can still be mapped and this causes issues at the moment as we'll emit a memcpy for function arguments to store to a local variable for certain edge cases, when we perform this memcpy on a null input, we cause a segfault at runtime. The fix to this is to simply create a branch around the store that checks if the data we're copying from is actually present. If it is, we proceed with the store, if it isn't we skip it. Full diff: https://github.com/llvm/llvm-project/pull/135477.diff 3 Files Affected:
diff --git a/flang/lib/Optimizer/OpenMP/MapInfoFinalization.cpp b/flang/lib/Optimizer/OpenMP/MapInfoFinalization.cpp
index 61f8713028a7f..3fcb4b04a7b76 100644
--- a/flang/lib/Optimizer/OpenMP/MapInfoFinalization.cpp
+++ b/flang/lib/Optimizer/OpenMP/MapInfoFinalization.cpp
@@ -153,7 +153,15 @@ class MapInfoFinalizationPass
builder.setInsertionPointToStart(allocaBlock);
auto alloca = builder.create<fir::AllocaOp>(loc, descriptor.getType());
builder.restoreInsertionPoint(insPt);
- builder.create<fir::StoreOp>(loc, descriptor, alloca);
+ // We should only emit a store if the passed in data is present, it is
+ // possible a user passes in no argument to an optional parameter, in which
+ // case we cannot store or we'll segfault on the emitted memcpy.
+ auto isPresent =
+ builder.create<fir::IsPresentOp>(loc, builder.getI1Type(), descriptor);
+ builder.genIfOp(loc, {}, isPresent, false)
+ .genThen(
+ [&]() { builder.create<fir::StoreOp>(loc, descriptor, alloca); })
+ .end();
return slot = alloca;
}
diff --git a/flang/test/Lower/OpenMP/optional-argument-map.f90 b/flang/test/Lower/OpenMP/optional-argument-map.f90
new file mode 100644
index 0000000000000..863742ce67927
--- /dev/null
+++ b/flang/test/Lower/OpenMP/optional-argument-map.f90
@@ -0,0 +1,27 @@
+!RUN: %flang_fc1 -emit-hlfir -fopenmp %s -o - | FileCheck %s
+
+module foo
+ implicit none
+ contains
+ subroutine test(I,A)
+ implicit none
+ real(4), optional, intent(inout) :: A(:)
+ integer(kind=4), intent(in) :: I
+
+ !$omp target data map(to: A) if (I>0)
+ !$omp end target data
+
+ end subroutine test
+end module foo
+
+! CHECK-LABEL: func.func @_QMfooPtest(
+! CHECK-SAME: %[[VAL_0:.*]]: !fir.ref<i32> {fir.bindc_name = "i"},
+! CHECK-SAME: %[[VAL_1:.*]]: !fir.box<!fir.array<?xf32>> {fir.bindc_name = "a", fir.optional}) {
+! CHECK: %[[VAL_2:.*]] = fir.alloca !fir.box<!fir.array<?xf32>>
+! CHECK: %[[VAL_3:.*]]:2 = hlfir.declare %[[VAL_1]] dummy_scope %{{.*}} {fortran_attrs = #fir.var_attrs<intent_inout, optional>, uniq_name = "_QMfooFtestEa"} : (!fir.box<!fir.array<?xf32>>, !fir.dscope) -> (!fir.box<!fir.array<?xf32>>, !fir.box<!fir.array<?xf32>>)
+! CHECK: %{{.*}} = fir.is_present %{{.*}}#1 : (!fir.box<!fir.array<?xf32>>) -> i1
+! CHECK: %{{.*}}:5 = fir.if %{{.*}}
+! CHECK: %[[VAL_4:.*]] = fir.is_present %[[VAL_3]]#1 : (!fir.box<!fir.array<?xf32>>) -> i1
+! CHECK: fir.if %[[VAL_4]] {
+! CHECK: fir.store %[[VAL_3]]#1 to %[[VAL_2]] : !fir.ref<!fir.box<!fir.array<?xf32>>>
+! CHECK: }
diff --git a/offload/test/offloading/fortran/optional-mapped-arguments.f90 b/offload/test/offloading/fortran/optional-mapped-arguments.f90
new file mode 100644
index 0000000000000..e0cdddb8e1bf7
--- /dev/null
+++ b/offload/test/offloading/fortran/optional-mapped-arguments.f90
@@ -0,0 +1,36 @@
+! OpenMP offloading test that checks we do not cause a segfault when mapping
+! optional function arguments (present or otherwise). No results requiring
+! checking other than that the program compiles and runs to completion with no
+! error.
+! REQUIRES: flang, amdgpu
+
+! RUN: %libomptarget-compile-fortran-run-and-check-generic
+module foo
+ implicit none
+ contains
+ subroutine test(I,A)
+ implicit none
+ real(4), optional, intent(inout) :: A(:)
+ integer(kind=4), intent(in) :: I
+
+ !$omp target data map(to: A) if (I>0)
+ !$omp end target data
+
+ !$omp target enter data map(to:A) if (I>0)
+
+ !$omp target exit data map(from:A) if (I>0)
+ end subroutine test
+end module foo
+
+program main
+ use foo
+ implicit none
+ real :: array(10)
+ call test(0)
+ call test(1)
+ call test(0, array)
+ call test(1, array)
+ print *, "PASSED"
+end program main
+
+! CHECK: PASSED
|
Member
|
@llvm/pr-subscribers-offload Author: None (agozillon) ChangesCurrently we don't check for the presence of descriptor/BoxTypes before emitting stores which lower to memcpys, the issue with this is that users can have optional arguments, where they don't provide an input, making the argument effectively null. This can still be mapped and this causes issues at the moment as we'll emit a memcpy for function arguments to store to a local variable for certain edge cases, when we perform this memcpy on a null input, we cause a segfault at runtime. The fix to this is to simply create a branch around the store that checks if the data we're copying from is actually present. If it is, we proceed with the store, if it isn't we skip it. Full diff: https://github.com/llvm/llvm-project/pull/135477.diff 3 Files Affected:
diff --git a/flang/lib/Optimizer/OpenMP/MapInfoFinalization.cpp b/flang/lib/Optimizer/OpenMP/MapInfoFinalization.cpp
index 61f8713028a7f..3fcb4b04a7b76 100644
--- a/flang/lib/Optimizer/OpenMP/MapInfoFinalization.cpp
+++ b/flang/lib/Optimizer/OpenMP/MapInfoFinalization.cpp
@@ -153,7 +153,15 @@ class MapInfoFinalizationPass
builder.setInsertionPointToStart(allocaBlock);
auto alloca = builder.create<fir::AllocaOp>(loc, descriptor.getType());
builder.restoreInsertionPoint(insPt);
- builder.create<fir::StoreOp>(loc, descriptor, alloca);
+ // We should only emit a store if the passed in data is present, it is
+ // possible a user passes in no argument to an optional parameter, in which
+ // case we cannot store or we'll segfault on the emitted memcpy.
+ auto isPresent =
+ builder.create<fir::IsPresentOp>(loc, builder.getI1Type(), descriptor);
+ builder.genIfOp(loc, {}, isPresent, false)
+ .genThen(
+ [&]() { builder.create<fir::StoreOp>(loc, descriptor, alloca); })
+ .end();
return slot = alloca;
}
diff --git a/flang/test/Lower/OpenMP/optional-argument-map.f90 b/flang/test/Lower/OpenMP/optional-argument-map.f90
new file mode 100644
index 0000000000000..863742ce67927
--- /dev/null
+++ b/flang/test/Lower/OpenMP/optional-argument-map.f90
@@ -0,0 +1,27 @@
+!RUN: %flang_fc1 -emit-hlfir -fopenmp %s -o - | FileCheck %s
+
+module foo
+ implicit none
+ contains
+ subroutine test(I,A)
+ implicit none
+ real(4), optional, intent(inout) :: A(:)
+ integer(kind=4), intent(in) :: I
+
+ !$omp target data map(to: A) if (I>0)
+ !$omp end target data
+
+ end subroutine test
+end module foo
+
+! CHECK-LABEL: func.func @_QMfooPtest(
+! CHECK-SAME: %[[VAL_0:.*]]: !fir.ref<i32> {fir.bindc_name = "i"},
+! CHECK-SAME: %[[VAL_1:.*]]: !fir.box<!fir.array<?xf32>> {fir.bindc_name = "a", fir.optional}) {
+! CHECK: %[[VAL_2:.*]] = fir.alloca !fir.box<!fir.array<?xf32>>
+! CHECK: %[[VAL_3:.*]]:2 = hlfir.declare %[[VAL_1]] dummy_scope %{{.*}} {fortran_attrs = #fir.var_attrs<intent_inout, optional>, uniq_name = "_QMfooFtestEa"} : (!fir.box<!fir.array<?xf32>>, !fir.dscope) -> (!fir.box<!fir.array<?xf32>>, !fir.box<!fir.array<?xf32>>)
+! CHECK: %{{.*}} = fir.is_present %{{.*}}#1 : (!fir.box<!fir.array<?xf32>>) -> i1
+! CHECK: %{{.*}}:5 = fir.if %{{.*}}
+! CHECK: %[[VAL_4:.*]] = fir.is_present %[[VAL_3]]#1 : (!fir.box<!fir.array<?xf32>>) -> i1
+! CHECK: fir.if %[[VAL_4]] {
+! CHECK: fir.store %[[VAL_3]]#1 to %[[VAL_2]] : !fir.ref<!fir.box<!fir.array<?xf32>>>
+! CHECK: }
diff --git a/offload/test/offloading/fortran/optional-mapped-arguments.f90 b/offload/test/offloading/fortran/optional-mapped-arguments.f90
new file mode 100644
index 0000000000000..e0cdddb8e1bf7
--- /dev/null
+++ b/offload/test/offloading/fortran/optional-mapped-arguments.f90
@@ -0,0 +1,36 @@
+! OpenMP offloading test that checks we do not cause a segfault when mapping
+! optional function arguments (present or otherwise). No results requiring
+! checking other than that the program compiles and runs to completion with no
+! error.
+! REQUIRES: flang, amdgpu
+
+! RUN: %libomptarget-compile-fortran-run-and-check-generic
+module foo
+ implicit none
+ contains
+ subroutine test(I,A)
+ implicit none
+ real(4), optional, intent(inout) :: A(:)
+ integer(kind=4), intent(in) :: I
+
+ !$omp target data map(to: A) if (I>0)
+ !$omp end target data
+
+ !$omp target enter data map(to:A) if (I>0)
+
+ !$omp target exit data map(from:A) if (I>0)
+ end subroutine test
+end module foo
+
+program main
+ use foo
+ implicit none
+ real :: array(10)
+ call test(0)
+ call test(1)
+ call test(0, array)
+ call test(1, array)
+ print *, "PASSED"
+end program main
+
+! CHECK: PASSED
|
Contributor
Author
|
Thank you very much for the quick review @skatrak :-) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Currently we don't check for the presence of descriptor/BoxTypes before emitting stores which lower to memcpys, the issue with this is that users can have optional arguments, where they don't provide an input, making the argument effectively null. This can still be mapped and this causes issues at the moment as we'll emit a memcpy for function arguments to store to a local variable for certain edge cases, when we perform this memcpy on a null input, we cause a segfault at runtime.
The fix to this is to simply create a branch around the store that checks if the data we're copying from is actually present. If it is, we proceed with the store, if it isn't we skip it.