[flang] Allow to pass an async id to allocate the descriptor (#118713)

This is a patch in preparation for the support stream ordered memory
allocator in CUDA Fortran.

This patch adds an asynchronous id to the AllocatableAllocate runtime
function and to Descriptor::Allocate so it can be passed down to the
registered allocator. It is up to the allocator to use this value or
not.

A follow up patch will implement that asynchronous allocator for CUDA
Fortran.

flang/include/flang/Runtime/CUDA/allocator.h

@@ -19,16 +19,16 @@ extern "C" {
 void RTDECL(CUFRegisterAllocator)();
 }
 
-void *CUFAllocPinned(std::size_t);
+void *CUFAllocPinned(std::size_t, std::int64_t);
 void CUFFreePinned(void *);
 
-void *CUFAllocDevice(std::size_t);
+void *CUFAllocDevice(std::size_t, std::int64_t);
 void CUFFreeDevice(void *);
 
-void *CUFAllocManaged(std::size_t);
+void *CUFAllocManaged(std::size_t, std::int64_t);
 void CUFFreeManaged(void *);
 
-void *CUFAllocUnified(std::size_t);
+void *CUFAllocUnified(std::size_t, std::int64_t);
 void CUFFreeUnified(void *);
 
 } // namespace Fortran::runtime::cuda

flang/include/flang/Runtime/CUDA/common.h

@@ -23,6 +23,9 @@ static constexpr unsigned kHostToDevice = 0;
 static constexpr unsigned kDeviceToHost = 1;
 static constexpr unsigned kDeviceToDevice = 2;
 
+/// Value used for asyncId when no specific stream is specified.
+static constexpr std::int64_t kCudaNoStream = -1;
+
 #define CUDA_REPORT_IF_ERROR(expr) \
   [](cudaError_t err) { \
     if (err == cudaSuccess) \

flang/include/flang/Runtime/allocatable.h

@@ -94,9 +94,9 @@ int RTDECL(AllocatableCheckLengthParameter)(Descriptor &,
 // Successfully allocated memory is initialized if the allocatable has a
 // derived type, and is always initialized by AllocatableAllocateSource().
 // Performs all necessary coarray synchronization and validation actions.
-int RTDECL(AllocatableAllocate)(Descriptor &, bool hasStat = false,
+int RTDECL(AllocatableAllocate)(Descriptor &, std::int64_t asyncId = -1,
-    const Descriptor *errMsg = nullptr, const char *sourceFile = nullptr,
+    bool hasStat = false, const Descriptor *errMsg = nullptr,
-    int sourceLine = 0);
+    const char *sourceFile = nullptr, int sourceLine = 0);
 int RTDECL(AllocatableAllocateSource)(Descriptor &, const Descriptor &source,
     bool hasStat = false, const Descriptor *errMsg = nullptr,
     const char *sourceFile = nullptr, int sourceLine = 0);

flang/include/flang/Runtime/allocator-registry.h

@@ -10,6 +10,7 @@
 #define FORTRAN_RUNTIME_ALLOCATOR_REGISTRY_H_
 
 #include "flang/Common/api-attrs.h"
+#include <cstdint>
 #include <cstdlib>
 #include <vector>
 
@@ -25,7 +26,7 @@ static constexpr unsigned kUnifiedAllocatorPos = 4;
 
 namespace Fortran::runtime {
 
-using AllocFct = void *(*)(std::size_t);
+using AllocFct = void *(*)(std::size_t, std::int64_t);
 using FreeFct = void (*)(void *);
 
 typedef struct Allocator_t {
@@ -33,10 +34,11 @@ typedef struct Allocator_t {
   FreeFct free{nullptr};
 } Allocator_t;
 
-#ifdef RT_DEVICE_COMPILATION
+static RT_API_ATTRS void *MallocWrapper(
-static RT_API_ATTRS void *MallocWrapper(std::size_t size) {
+    std::size_t size, [[maybe_unused]] std::int64_t) {
   return std::malloc(size);
 }
+#ifdef RT_DEVICE_COMPILATION
 static RT_API_ATTRS void FreeWrapper(void *p) { return std::free(p); }
 #endif
 
@@ -46,7 +48,7 @@ struct AllocatorRegistry {
       : allocators{{&MallocWrapper, &FreeWrapper}} {}
 #else
   constexpr AllocatorRegistry() {
-    allocators[kDefaultAllocator] = {&std::malloc, &std::free};
+    allocators[kDefaultAllocator] = {&MallocWrapper, &std::free};
   };
 #endif
   RT_API_ATTRS void Register(int, Allocator_t);

flang/include/flang/Runtime/descriptor.h

@@ -374,7 +374,7 @@ public:
   // before calling.  It (re)computes the byte strides after
   // allocation.  Does not allocate automatic components or
   // perform default component initialization.
-  RT_API_ATTRS int Allocate();
+  RT_API_ATTRS int Allocate(std::int64_t asyncId = -1);
   RT_API_ATTRS void SetByteStrides();
 
   // Deallocates storage; does not call FINAL subroutines or

flang/lib/Lower/Allocatable.cpp

@@ -184,9 +184,14 @@ static mlir::Value genRuntimeAllocate(fir::FirOpBuilder &builder,
           ? fir::runtime::getRuntimeFunc<mkRTKey(PointerAllocate)>(loc, builder)
           : fir::runtime::getRuntimeFunc<mkRTKey(AllocatableAllocate)>(loc,
                                                                        builder);
-  llvm::SmallVector<mlir::Value> args{
+  llvm::SmallVector<mlir::Value> args{box.getAddr()};
-      box.getAddr(), errorManager.hasStat, errorManager.errMsgAddr,
+  if (!box.isPointer())
-      errorManager.sourceFile, errorManager.sourceLine};
+    args.push_back(
+        builder.createIntegerConstant(loc, builder.getI64Type(), -1));
+  args.push_back(errorManager.hasStat);
+  args.push_back(errorManager.errMsgAddr);
+  args.push_back(errorManager.sourceFile);
+  args.push_back(errorManager.sourceLine);
   llvm::SmallVector<mlir::Value> operands;
   for (auto [fst, snd] : llvm::zip(args, callee.getFunctionType().getInputs()))
     operands.emplace_back(builder.createConvert(loc, snd, fst));

flang/lib/Optimizer/Builder/Runtime/Allocatable.cpp

@@ -76,16 +76,19 @@ void fir::runtime::genAllocatableAllocate(fir::FirOpBuilder &builder,
   mlir::func::FuncOp func{
       fir::runtime::getRuntimeFunc<mkRTKey(AllocatableAllocate)>(loc, builder)};
   mlir::FunctionType fTy{func.getFunctionType()};
+  mlir::Value asyncId =
+      builder.createIntegerConstant(loc, builder.getI64Type(), -1);
   mlir::Value sourceFile{fir::factory::locationToFilename(builder, loc)};
   mlir::Value sourceLine{
-      fir::factory::locationToLineNo(builder, loc, fTy.getInput(4))};
+      fir::factory::locationToLineNo(builder, loc, fTy.getInput(5))};
   if (!hasStat)
     hasStat = builder.createBool(loc, false);
   if (!errMsg) {
     mlir::Type boxNoneTy = fir::BoxType::get(builder.getNoneType());
     errMsg = builder.create<fir::AbsentOp>(loc, boxNoneTy).getResult();
   }
-  llvm::SmallVector<mlir::Value> args{fir::runtime::createArguments(
+  llvm::SmallVector<mlir::Value> args{
-      builder, loc, fTy, desc, hasStat, errMsg, sourceFile, sourceLine)};
+      fir::runtime::createArguments(builder, loc, fTy, desc, asyncId, hasStat,
+                                    errMsg, sourceFile, sourceLine)};
   builder.create<fir::CallOp>(loc, func, args);
 }

flang/runtime/CUDA/allocatable.cpp

@@ -52,7 +52,7 @@ int RTDEF(CUFAllocatableAllocate)(Descriptor &desc, int64_t stream,
   }
   // Perform the standard allocation.
   int stat{RTNAME(AllocatableAllocate)(
-      desc, hasStat, errMsg, sourceFile, sourceLine)};
+      desc, stream, hasStat, errMsg, sourceFile, sourceLine)};
   return stat;
 }
 

flang/runtime/CUDA/allocator.cpp

@@ -33,7 +33,8 @@ void RTDEF(CUFRegisterAllocator)() {
 }
 }
 
-void *CUFAllocPinned(std::size_t sizeInBytes) {
+void *CUFAllocPinned(
+    std::size_t sizeInBytes, [[maybe_unused]] std::int64_t asyncId) {
   void *p;
   CUDA_REPORT_IF_ERROR(cudaMallocHost((void **)&p, sizeInBytes));
   return p;
@@ -41,7 +42,8 @@ void *CUFAllocPinned(std::size_t sizeInBytes) {
 
 void CUFFreePinned(void *p) { CUDA_REPORT_IF_ERROR(cudaFreeHost(p)); }
 
-void *CUFAllocDevice(std::size_t sizeInBytes) {
+void *CUFAllocDevice(
+    std::size_t sizeInBytes, [[maybe_unused]] std::int64_t asyncId) {
   void *p;
   CUDA_REPORT_IF_ERROR(cudaMalloc(&p, sizeInBytes));
   return p;
@@ -49,7 +51,8 @@ void *CUFAllocDevice(std::size_t sizeInBytes) {
 
 void CUFFreeDevice(void *p) { CUDA_REPORT_IF_ERROR(cudaFree(p)); }
 
-void *CUFAllocManaged(std::size_t sizeInBytes) {
+void *CUFAllocManaged(
+    std::size_t sizeInBytes, [[maybe_unused]] std::int64_t asyncId) {
   void *p;
   CUDA_REPORT_IF_ERROR(
       cudaMallocManaged((void **)&p, sizeInBytes, cudaMemAttachGlobal));
@@ -58,9 +61,10 @@ void *CUFAllocManaged(std::size_t sizeInBytes) {
 
 void CUFFreeManaged(void *p) { CUDA_REPORT_IF_ERROR(cudaFree(p)); }
 
-void *CUFAllocUnified(std::size_t sizeInBytes) {
+void *CUFAllocUnified(
+    std::size_t sizeInBytes, [[maybe_unused]] std::int64_t asyncId) {
   // Call alloc managed for the time being.
-  return CUFAllocManaged(sizeInBytes);
+  return CUFAllocManaged(sizeInBytes, asyncId);
 }
 
 void CUFFreeUnified(void *p) {

flang/runtime/CUDA/descriptor.cpp

@@ -19,7 +19,8 @@ RT_EXT_API_GROUP_BEGIN
 
 Descriptor *RTDEF(CUFAllocDesciptor)(
     std::size_t sizeInBytes, const char *sourceFile, int sourceLine) {
-  return reinterpret_cast<Descriptor *>(CUFAllocManaged(sizeInBytes));
+  return reinterpret_cast<Descriptor *>(
+      CUFAllocManaged(sizeInBytes, kCudaNoStream));
 }
 
 void RTDEF(CUFFreeDesciptor)(

flang/runtime/allocatable.cpp

@@ -133,15 +133,17 @@ void RTDEF(AllocatableApplyMold)(
   }
 }
 
-int RTDEF(AllocatableAllocate)(Descriptor &descriptor, bool hasStat,
+int RTDEF(AllocatableAllocate)(Descriptor &descriptor, std::int64_t asyncId,
-    const Descriptor *errMsg, const char *sourceFile, int sourceLine) {
+    bool hasStat, const Descriptor *errMsg, const char *sourceFile,
+    int sourceLine) {
   Terminator terminator{sourceFile, sourceLine};
   if (!descriptor.IsAllocatable()) {
     return ReturnError(terminator, StatInvalidDescriptor, errMsg, hasStat);
   } else if (descriptor.IsAllocated()) {
     return ReturnError(terminator, StatBaseNotNull, errMsg, hasStat);
   } else {
-    int stat{ReturnError(terminator, descriptor.Allocate(), errMsg, hasStat)};
+    int stat{
+        ReturnError(terminator, descriptor.Allocate(asyncId), errMsg, hasStat)};
     if (stat == StatOk) {
       if (const DescriptorAddendum * addendum{descriptor.Addendum()}) {
         if (const auto *derived{addendum->derivedType()}) {
@@ -160,7 +162,7 @@ int RTDEF(AllocatableAllocateSource)(Descriptor &alloc,
     const Descriptor &source, bool hasStat, const Descriptor *errMsg,
     const char *sourceFile, int sourceLine) {
   int stat{RTNAME(AllocatableAllocate)(
-      alloc, hasStat, errMsg, sourceFile, sourceLine)};
+      alloc, /*asyncId=*/-1, hasStat, errMsg, sourceFile, sourceLine)};
   if (stat == StatOk) {
     Terminator terminator{sourceFile, sourceLine};
     DoFromSourceAssign(alloc, source, terminator);

flang/runtime/array-constructor.cpp

@@ -50,8 +50,8 @@ static RT_API_ATTRS void AllocateOrReallocateVectorIfNeeded(
           initialAllocationSize(fromElements, to.ElementBytes())};
       to.GetDimension(0).SetBounds(1, allocationSize);
       RTNAME(AllocatableAllocate)
-      (to, /*hasStat=*/false, /*errMsg=*/nullptr, vector.sourceFile,
+      (to, /*asyncId=*/-1, /*hasStat=*/false, /*errMsg=*/nullptr,
-          vector.sourceLine);
+          vector.sourceFile, vector.sourceLine);
       to.GetDimension(0).SetBounds(1, fromElements);
       vector.actualAllocationSize = allocationSize;
     } else {
@@ -59,8 +59,8 @@ static RT_API_ATTRS void AllocateOrReallocateVectorIfNeeded(
       // first value: there should be no reallocation.
       RUNTIME_CHECK(terminator, previousToElements >= fromElements);
       RTNAME(AllocatableAllocate)
-      (to, /*hasStat=*/false, /*errMsg=*/nullptr, vector.sourceFile,
+      (to, /*asyncId=*/-1, /*hasStat=*/false, /*errMsg=*/nullptr,
-          vector.sourceLine);
+          vector.sourceFile, vector.sourceLine);
       vector.actualAllocationSize = previousToElements;
     }
   } else {

flang/runtime/descriptor.cpp

@@ -163,7 +163,7 @@ RT_API_ATTRS static inline int MapAllocIdx(const Descriptor &desc) {
 #endif
 }
 
-RT_API_ATTRS int Descriptor::Allocate() {
+RT_API_ATTRS int Descriptor::Allocate(std::int64_t asyncId) {
   std::size_t elementBytes{ElementBytes()};
   if (static_cast<std::int64_t>(elementBytes) < 0) {
     // F'2023 7.4.4.2 p5: "If the character length parameter value evaluates
@@ -175,7 +175,7 @@ RT_API_ATTRS int Descriptor::Allocate() {
   // Zero size allocation is possible in Fortran and the resulting
   // descriptor must be allocated/associated. Since std::malloc(0)
   // result is implementation defined, always allocate at least one byte.
-  void *p{alloc(byteSize ? byteSize : 1)};
+  void *p{alloc(byteSize ? byteSize : 1, asyncId)};
   if (!p) {
     return CFI_ERROR_MEM_ALLOCATION;
   }

flang/test/HLFIR/elemental-codegen.fir

@@ -192,7 +192,7 @@ func.func @test_polymorphic(%arg0: !fir.class<!fir.type<_QMtypesTt>> {fir.bindc_
 // CHECK:           %[[VAL_35:.*]] = fir.absent !fir.box<none>
 // CHECK:           %[[VAL_36:.*]] = fir.convert %[[VAL_4]] : (!fir.ref<!fir.class<!fir.heap<!fir.array<?x?x!fir.type<_QMtypesTt>>>>>) -> !fir.ref<!fir.box<none>>
 // CHECK:           %[[VAL_37:.*]] = fir.convert %[[VAL_31]] : (!fir.ref<!fir.char<1,{{.*}}>>) -> !fir.ref<i8>
-// CHECK:           %[[VAL_38:.*]] = fir.call @_FortranAAllocatableAllocate(%[[VAL_36]], %[[VAL_34]], %[[VAL_35]], %[[VAL_37]], %[[VAL_33]]) : (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+// CHECK:           %[[VAL_38:.*]] = fir.call @_FortranAAllocatableAllocate(%[[VAL_36]], %{{.*}}, %[[VAL_34]], %[[VAL_35]], %[[VAL_37]], %[[VAL_33]]) : (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 // CHECK:           %[[VAL_39:.*]] = fir.load %[[VAL_13]]#0 : !fir.ref<!fir.class<!fir.heap<!fir.array<?x?x!fir.type<_QMtypesTt>>>>>
 // CHECK:           %[[VAL_40:.*]] = arith.constant 1 : index
 // CHECK:           fir.do_loop %[[VAL_41:.*]] = %[[VAL_40]] to %[[EX1]] step %[[VAL_40]] unordered {
@@ -276,7 +276,7 @@ func.func @test_polymorphic_expr(%arg0: !fir.class<!fir.type<_QMtypesTt>> {fir.b
 // CHECK:           %[[VAL_36:.*]] = fir.absent !fir.box<none>
 // CHECK:           %[[VAL_37:.*]] = fir.convert %[[VAL_5]] : (!fir.ref<!fir.class<!fir.heap<!fir.array<?x?x!fir.type<_QMtypesTt>>>>>) -> !fir.ref<!fir.box<none>>
 // CHECK:           %[[VAL_38:.*]] = fir.convert %[[VAL_32]] : (!fir.ref<!fir.char<1,{{.*}}>>) -> !fir.ref<i8>
-// CHECK:           %[[VAL_39:.*]] = fir.call @_FortranAAllocatableAllocate(%[[VAL_37]], %[[VAL_35]], %[[VAL_36]], %[[VAL_38]], %[[VAL_34]]) : (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+// CHECK:           %[[VAL_39:.*]] = fir.call @_FortranAAllocatableAllocate(%[[VAL_37]], %{{.*}}, %[[VAL_35]], %[[VAL_36]], %[[VAL_38]], %[[VAL_34]]) : (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 // CHECK:           %[[VAL_40:.*]] = fir.load %[[VAL_14]]#0 : !fir.ref<!fir.class<!fir.heap<!fir.array<?x?x!fir.type<_QMtypesTt>>>>>
 // CHECK:           %[[VAL_41:.*]] = arith.constant 1 : index
 // CHECK:           fir.do_loop %[[VAL_42:.*]] = %[[VAL_41]] to %[[VAL_3]] step %[[VAL_41]] unordered {
@@ -329,7 +329,7 @@ func.func @test_polymorphic_expr(%arg0: !fir.class<!fir.type<_QMtypesTt>> {fir.b
 // CHECK:           %[[VAL_85:.*]] = fir.absent !fir.box<none>
 // CHECK:           %[[VAL_86:.*]] = fir.convert %[[VAL_4]] : (!fir.ref<!fir.class<!fir.heap<!fir.array<?x?x!fir.type<_QMtypesTt>>>>>) -> !fir.ref<!fir.box<none>>
 // CHECK:           %[[VAL_87:.*]] = fir.convert %[[VAL_81]] : (!fir.ref<!fir.char<1,{{.*}}>>) -> !fir.ref<i8>
-// CHECK:           %[[VAL_88:.*]] = fir.call @_FortranAAllocatableAllocate(%[[VAL_86]], %[[VAL_84]], %[[VAL_85]], %[[VAL_87]], %[[VAL_83]]) : (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+// CHECK:           %[[VAL_88:.*]] = fir.call @_FortranAAllocatableAllocate(%[[VAL_86]], %{{.*}}, %[[VAL_84]], %[[VAL_85]], %[[VAL_87]], %[[VAL_83]]) : (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 // CHECK:           %[[VAL_89:.*]] = fir.load %[[VAL_63]]#0 : !fir.ref<!fir.class<!fir.heap<!fir.array<?x?x!fir.type<_QMtypesTt>>>>>
 // CHECK:           %[[VAL_90:.*]] = arith.constant 1 : index
 // CHECK:           fir.do_loop %[[VAL_91:.*]] = %[[VAL_90]] to %[[VAL_3]] step %[[VAL_90]] unordered {

flang/test/Lower/OpenACC/acc-declare.f90

@@ -469,6 +469,6 @@ contains
 end module
 
 ! CHECK-LABEL: func.func @_QMacc_declare_post_action_statPinit()
-! CHECK: fir.call @_FortranAAllocatableAllocate({{.*}}) fastmath<contract> {acc.declare_action = #acc.declare_action<postAlloc = @_QMacc_declare_post_action_statEx_acc_declare_update_desc_post_alloc>} : (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: fir.call @_FortranAAllocatableAllocate({{.*}}) fastmath<contract> {acc.declare_action = #acc.declare_action<postAlloc = @_QMacc_declare_post_action_statEx_acc_declare_update_desc_post_alloc>} : (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 ! CHECK: fir.if
-! CHECK: fir.call @_FortranAAllocatableAllocate({{.*}}) fastmath<contract> {acc.declare_action = #acc.declare_action<postAlloc = @_QMacc_declare_post_action_statEy_acc_declare_update_desc_post_alloc>} : (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: fir.call @_FortranAAllocatableAllocate({{.*}}) fastmath<contract> {acc.declare_action = #acc.declare_action<postAlloc = @_QMacc_declare_post_action_statEy_acc_declare_update_desc_post_alloc>} : (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32

flang/test/Lower/allocatable-polymorphic.f90

@@ -267,7 +267,7 @@ contains
 ! CHECK: %[[C0:.*]] = arith.constant 0 : i32
 ! CHECK: fir.call @_FortranAAllocatableInitDerivedForAllocate(%[[P_CAST]], %[[TYPE_DESC_P1_CAST]], %[[RANK]], %[[C0]]) {{.*}}: (!fir.ref<!fir.box<none>>, !fir.ref<none>, i32, i32) -> none
 ! CHECK: %[[P_CAST:.*]] = fir.convert %[[P_DECL]]#1 : (!fir.ref<!fir.class<!fir.heap<!fir.type<_QMpolyTp1{a:i32,b:i32}>>>>) -> !fir.ref<!fir.box<none>>
-! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[P_CAST]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}}: (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[P_CAST]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}}: (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 
 ! CHECK: %[[TYPE_DESC_P1:.*]] = fir.type_desc !fir.type<_QMpolyTp1{a:i32,b:i32}>
 ! CHECK: %[[C1_CAST:.*]] = fir.convert %[[C1_DECL]]#1 : (!fir.ref<!fir.class<!fir.heap<!fir.type<_QMpolyTp1{a:i32,b:i32}>>>>) -> !fir.ref<!fir.box<none>>
@@ -276,7 +276,7 @@ contains
 ! CHECK: %[[C0:.*]] = arith.constant 0 : i32
 ! CHECK: fir.call @_FortranAAllocatableInitDerivedForAllocate(%[[C1_CAST]], %[[TYPE_DESC_P1_CAST]], %[[RANK]], %[[C0]]) {{.*}}: (!fir.ref<!fir.box<none>>, !fir.ref<none>, i32, i32) -> none
 ! CHECK: %[[C1_CAST:.*]] = fir.convert %[[C1_DECL]]#1 : (!fir.ref<!fir.class<!fir.heap<!fir.type<_QMpolyTp1{a:i32,b:i32}>>>>) -> !fir.ref<!fir.box<none>>
-! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[C1_CAST]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}}: (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[C1_CAST]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}}: (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 
 ! CHECK: %[[TYPE_DESC_P2:.*]] = fir.type_desc !fir.type<_QMpolyTp2{p1:!fir.type<_QMpolyTp1{a:i32,b:i32}>,c:i32}>
 ! CHECK: %[[C2_CAST:.*]] = fir.convert %[[C2_DECL]]#1 : (!fir.ref<!fir.class<!fir.heap<!fir.type<_QMpolyTp1{a:i32,b:i32}>>>>) -> !fir.ref<!fir.box<none>>
@@ -285,7 +285,7 @@ contains
 ! CHECK: %[[C0:.*]] = arith.constant 0 : i32
 ! CHECK: fir.call @_FortranAAllocatableInitDerivedForAllocate(%[[C2_CAST]], %[[TYPE_DESC_P2_CAST]], %[[RANK]], %[[C0]]) {{.*}}: (!fir.ref<!fir.box<none>>, !fir.ref<none>, i32, i32) -> none
 ! CHECK: %[[C2_CAST:.*]] = fir.convert %[[C2_DECL]]#1 : (!fir.ref<!fir.class<!fir.heap<!fir.type<_QMpolyTp1{a:i32,b:i32}>>>>) -> !fir.ref<!fir.box<none>>
-! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[C2_CAST]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}}: (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[C2_CAST]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}}: (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 
 ! CHECK: %[[TYPE_DESC_P1:.*]] = fir.type_desc !fir.type<_QMpolyTp1{a:i32,b:i32}>
 ! CHECK: %[[C3_CAST:.*]] = fir.convert %[[C3_DECL]]#1 : (!fir.ref<!fir.class<!fir.heap<!fir.array<?x!fir.type<_QMpolyTp1{a:i32,b:i32}>>>>>) -> !fir.ref<!fir.box<none>>
@@ -300,7 +300,7 @@ contains
 ! CHECK: %[[C10_I64:.*]] = fir.convert %[[C10]] : (i32) -> i64
 ! CHECK: %{{.*}} = fir.call @_FortranAAllocatableSetBounds(%[[C3_CAST]], %[[C0]], %[[C1_I64]], %[[C10_I64]]) {{.*}}: (!fir.ref<!fir.box<none>>, i32, i64, i64) -> none
 ! CHECK: %[[C3_CAST:.*]] = fir.convert %[[C3_DECL]]#1 : (!fir.ref<!fir.class<!fir.heap<!fir.array<?x!fir.type<_QMpolyTp1{a:i32,b:i32}>>>>>) -> !fir.ref<!fir.box<none>>
-! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[C3_CAST]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}}: (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[C3_CAST]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}}: (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 
 ! CHECK: %[[TYPE_DESC_P2:.*]] = fir.type_desc !fir.type<_QMpolyTp2{p1:!fir.type<_QMpolyTp1{a:i32,b:i32}>,c:i32}>
 ! CHECK: %[[C4_CAST:.*]] = fir.convert %[[C4_DECL]]#1 : (!fir.ref<!fir.class<!fir.heap<!fir.array<?x!fir.type<_QMpolyTp1{a:i32,b:i32}>>>>>) -> !fir.ref<!fir.box<none>>
@@ -316,7 +316,7 @@ contains
 ! CHECK: %[[C20_I64:.*]] = fir.convert %[[C20]] : (i32) -> i64
 ! CHECK: %{{.*}} = fir.call @_FortranAAllocatableSetBounds(%[[C4_CAST]], %[[C0]], %[[C1_I64]], %[[C20_I64]]) {{.*}}: (!fir.ref<!fir.box<none>>, i32, i64, i64) -> none
 ! CHECK: %[[C4_CAST:.*]] = fir.convert %[[C4_DECL]]#1 : (!fir.ref<!fir.class<!fir.heap<!fir.array<?x!fir.type<_QMpolyTp1{a:i32,b:i32}>>>>>) -> !fir.ref<!fir.box<none>>
-! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[C4_CAST]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}}: (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[C4_CAST]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}}: (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 
 ! CHECK: %[[C1_LOAD1:.*]] = fir.load %[[C1_DECL]]#0 : !fir.ref<!fir.class<!fir.heap<!fir.type<_QMpolyTp1{a:i32,b:i32}>>>>
 ! CHECK: fir.dispatch "proc1"(%[[C1_LOAD1]] : !fir.class<!fir.heap<!fir.type<_QMpolyTp1{a:i32,b:i32}>>>)
@@ -390,7 +390,7 @@ contains
 ! CHECK: %[[CORANK:.*]] = arith.constant 0 : i32
 ! CHECK: %{{.*}} = fir.call @_FortranAAllocatableInitIntrinsicForAllocate(%[[BOX_NONE]], %[[CAT]], %[[KIND]], %[[RANK]], %[[CORANK]]) {{.*}} : (!fir.ref<!fir.box<none>>, i32, i32, i32, i32) -> none
 ! CHECK: %[[BOX_NONE:.*]] = fir.convert %[[P_DECL]]#1 : (!fir.ref<!fir.class<!fir.heap<none>>>) -> !fir.ref<!fir.box<none>>
-! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[BOX_NONE]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[BOX_NONE]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 
 ! CHECK: %[[BOX_NONE:.*]] = fir.convert %[[PTR_DECL]]#1 : (!fir.ref<!fir.class<!fir.ptr<none>>>) -> !fir.ref<!fir.box<none>>
 ! CHECK: %[[CAT:.*]] = arith.constant 1 : i32
@@ -573,7 +573,7 @@ contains
 ! CHECK: %[[CORANK:.*]] = arith.constant 0 : i32
 ! CHECK: %{{.*}} = fir.call @_FortranAAllocatableInitCharacterForAllocate(%[[A_NONE]], %[[LEN]], %[[KIND]], %[[RANK]], %[[CORANK]]) {{.*}} : (!fir.ref<!fir.box<none>>, i64, i32, i32, i32) -> none
 ! CHECK: %[[A_NONE:.*]] = fir.convert %[[A_DECL]]#1 : (!fir.ref<!fir.class<!fir.heap<none>>>) -> !fir.ref<!fir.box<none>>
-! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[A_NONE]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[A_NONE]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 
 end module
 
@@ -592,17 +592,17 @@ end
 ! LLVM-LABEL: define void @_QMpolyPtest_allocatable()
 
 ! LLVM: %{{.*}} = call {} @_FortranAAllocatableInitDerivedForAllocate(ptr %{{.*}}, ptr @_QMpolyEXdtXp1, i32 0, i32 0)
-! LLVM: %{{.*}} = call i32 @_FortranAAllocatableAllocate(ptr %{{.*}}, i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})
+! LLVM: %{{.*}} = call i32 @_FortranAAllocatableAllocate(ptr %{{.*}}, i64 -1, i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})
 ! LLVM: %{{.*}} = call {} @_FortranAAllocatableInitDerivedForAllocate(ptr %{{.*}}, ptr @_QMpolyEXdtXp1, i32 0, i32 0)
-! LLVM: %{{.*}} = call i32 @_FortranAAllocatableAllocate(ptr %{{.*}}, i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})
+! LLVM: %{{.*}} = call i32 @_FortranAAllocatableAllocate(ptr %{{.*}}, i64 -1, i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})
 ! LLVM: %{{.*}} = call {} @_FortranAAllocatableInitDerivedForAllocate(ptr %{{.*}}, ptr @_QMpolyEXdtXp2, i32 0, i32 0)
-! LLVM: %{{.*}} = call i32 @_FortranAAllocatableAllocate(ptr %{{.*}}, i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})
+! LLVM: %{{.*}} = call i32 @_FortranAAllocatableAllocate(ptr %{{.*}}, i64 -1, i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})
 ! LLVM: %{{.*}} = call {} @_FortranAAllocatableInitDerivedForAllocate(ptr %{{.*}}, ptr @_QMpolyEXdtXp1, i32 1, i32 0)
 ! LLVM: %{{.*}} = call {} @_FortranAAllocatableSetBounds(ptr %{{.*}}, i32 0, i64 1, i64 10)
-! LLVM: %{{.*}} = call i32 @_FortranAAllocatableAllocate(ptr %{{.*}}, i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})
+! LLVM: %{{.*}} = call i32 @_FortranAAllocatableAllocate(ptr %{{.*}}, i64 -1, i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})
 ! LLVM: %{{.*}} = call {} @_FortranAAllocatableInitDerivedForAllocate(ptr %{{.*}}, ptr @_QMpolyEXdtXp2, i32 1, i32 0)
 ! LLVM: %{{.*}} = call {} @_FortranAAllocatableSetBounds(ptr %{{.*}}, i32 0, i64 1, i64 20)
-! LLVM: %{{.*}} = call i32 @_FortranAAllocatableAllocate(ptr %{{.*}}, i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})
+! LLVM: %{{.*}} = call i32 @_FortranAAllocatableAllocate(ptr %{{.*}}, i64 -1, i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})
 ! LLVM-COUNT-2:  call void %{{[0-9]*}}()
 
 ! LLVM: call void @llvm.memcpy.p0.p0.i32
@@ -683,5 +683,5 @@ end
 ! LLVM: store { ptr, i64, i32, i8, i8, i8, i8, ptr, [1 x i64] } { ptr null, i64 ptrtoint (ptr getelementptr (%_QMpolyTp1, ptr null, i32 1) to i64), i32 20240719, i8 0, i8 42, i8 2, i8 1, ptr @_QMpolyEXdtXp1, [1 x i64] zeroinitializer }, ptr %[[ALLOCA1:[0-9]*]]
 ! LLVM: call void @llvm.memcpy.p0.p0.i32(ptr %[[ALLOCA2:[0-9]+]], ptr %[[ALLOCA1]], i32 40, i1 false)
 ! LLVM: %{{.*}} = call {} @_FortranAAllocatableInitDerivedForAllocate(ptr %[[ALLOCA2]], ptr @_QMpolyEXdtXp1, i32 0, i32 0)
-! LLVM: %{{.*}} = call i32 @_FortranAAllocatableAllocate(ptr %[[ALLOCA2]], i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})
+! LLVM: %{{.*}} = call i32 @_FortranAAllocatableAllocate(ptr %[[ALLOCA2]], i64 -1, i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})
 ! LLVM: %{{.*}} = call i32 @_FortranAAllocatableDeallocatePolymorphic(ptr %[[ALLOCA2]], ptr {{.*}}, i1 false, ptr null, ptr @_QQclX{{.*}}, i32 {{.*}})

flang/test/Lower/allocatable-runtime.f90

@@ -31,7 +31,7 @@ subroutine foo()
   ! CHECK: fir.call @{{.*}}AllocatableSetBounds(%[[xBoxCast2]], %c0{{.*}}, %[[xlbCast]], %[[xubCast]]) {{.*}}: (!fir.ref<!fir.box<none>>, i32, i64, i64) -> none
   ! CHECK-DAG: %[[xBoxCast3:.*]] = fir.convert %[[xBoxAddr]] : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> !fir.ref<!fir.box<none>>
   ! CHECK-DAG: %[[sourceFile:.*]] = fir.convert %{{.*}} -> !fir.ref<i8>
-  ! CHECK: fir.call @{{.*}}AllocatableAllocate(%[[xBoxCast3]], %false{{.*}}, %[[errMsg]], %[[sourceFile]], %{{.*}}) {{.*}}: (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+  ! CHECK: fir.call @{{.*}}AllocatableAllocate(%[[xBoxCast3]], %c-1{{.*}}, %false{{.*}}, %[[errMsg]], %[[sourceFile]], %{{.*}}) {{.*}}: (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 
   ! Simply check that we are emitting the right numebr of set bound for y and z. Otherwise, this is just like x.
   ! CHECK: fir.convert %[[yBoxAddr]] : (!fir.ref<!fir.box<!fir.heap<!fir.array<?x?xf32>>>>) -> !fir.ref<!fir.box<none>>
@@ -180,4 +180,4 @@ end subroutine
 ! CHECK: %[[M_BOX_NONE:.*]] = fir.convert %[[EMBOX_M]] : (!fir.box<!fir.array<10xi32>>) -> !fir.box<none>
 ! CHECK: %{{.*}} = fir.call @_FortranAAllocatableApplyMold(%[[A_BOX_NONE]], %[[M_BOX_NONE]], %[[RANK]]) {{.*}} : (!fir.ref<!fir.box<none>>, !fir.box<none>, i32) -> none
 ! CHECK: %[[A_BOX_NONE:.*]] = fir.convert %[[A]] : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xi32>>>>) -> !fir.ref<!fir.box<none>>
-! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[A_BOX_NONE]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[A_BOX_NONE]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32

flang/test/Lower/allocate-mold.f90

@@ -16,7 +16,7 @@ end subroutine
 ! CHECK: %[[A_REF_BOX_NONE1:.*]] = fir.convert %[[A]] : (!fir.ref<!fir.box<!fir.heap<i32>>>) -> !fir.ref<!fir.box<none>>
 ! CHECK: %{{.*}} = fir.call @_FortranAAllocatableApplyMold(%[[A_REF_BOX_NONE1]], %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, !fir.box<none>, i32) -> none
 ! CHECK: %[[A_REF_BOX_NONE2:.*]] = fir.convert %[[A]] : (!fir.ref<!fir.box<!fir.heap<i32>>>) -> !fir.ref<!fir.box<none>>
-! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[A_REF_BOX_NONE2]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[A_REF_BOX_NONE2]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 
 subroutine array_scalar_mold_allocation()
   real, allocatable :: a(:)
@@ -40,4 +40,4 @@ end subroutine array_scalar_mold_allocation
 ! CHECK: %[[REF_BOX_A1:.*]] = fir.convert %1 : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> !fir.ref<!fir.box<none>>
 ! CHECK: %{{.*}} = fir.call @_FortranAAllocatableSetBounds(%[[REF_BOX_A1]], {{.*}},{{.*}}, {{.*}}) fastmath<contract> : (!fir.ref<!fir.box<none>>, i32, i64, i64) -> none
 ! CHECK: %[[REF_BOX_A2:.*]] = fir.convert %[[A]] : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> !fir.ref<!fir.box<none>>
-! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[REF_BOX_A2]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[REF_BOX_A2]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32

flang/test/Lower/polymorphic.f90

@@ -1154,11 +1154,11 @@ end program
 ! CHECK-LABEL: func.func @_QQmain() attributes {fir.bindc_name = "test"} {
 ! CHECK: %[[ADDR_O:.*]] = fir.address_of(@_QFEo) : !fir.ref<!fir.box<!fir.heap<!fir.type<_QMpolymorphic_testTouter{inner:!fir.type<_QMpolymorphic_testTp1{a:i32,b:i32}>}>>>>
 ! CHECK: %[[BOX_NONE:.*]] = fir.convert %[[ADDR_O]] : (!fir.ref<!fir.box<!fir.heap<!fir.type<_QMpolymorphic_testTouter{inner:!fir.type<_QMpolymorphic_testTp1{a:i32,b:i32}>}>>>>) -> !fir.ref<!fir.box<none>>
-! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[BOX_NONE]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
+! CHECK: %{{.*}} = fir.call @_FortranAAllocatableAllocate(%[[BOX_NONE]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}}) {{.*}} : (!fir.ref<!fir.box<none>>, i64, i1, !fir.box<none>, !fir.ref<i8>, i32) -> i32
 ! CHECK: %[[O:.*]] = fir.load %[[ADDR_O]] : !fir.ref<!fir.box<!fir.heap<!fir.type<_QMpolymorphic_testTouter{inner:!fir.type<_QMpolymorphic_testTp1{a:i32,b:i32}>}>>>>
 ! CHECK: %[[FIELD_INNER:.*]] = fir.field_index inner, !fir.type<_QMpolymorphic_testTouter{inner:!fir.type<_QMpolymorphic_testTp1{a:i32,b:i32}>}>
 ! CHECK: %[[COORD_INNER:.*]] = fir.coordinate_of %[[O]], %[[FIELD_INNER]] : (!fir.box<!fir.heap<!fir.type<_QMpolymorphic_testTouter{inner:!fir.type<_QMpolymorphic_testTp1{a:i32,b:i32}>}>>>, !fir.field) -> !fir.ref<!fir.type<_QMpolymorphic_testTp1{a:i32,b:i32}>>
-! CHECK: %{{.*}} = fir.do_loop %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} unordered iter_args(%arg1 = %9) -> (!fir.array<5x!fir.logical<4>>) {
+! CHECK: %{{.*}} = fir.do_loop %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} unordered iter_args(%arg1 = %{{.*}}) -> (!fir.array<5x!fir.logical<4>>) {
 ! CHECK:   %[[EMBOXED:.*]] = fir.embox %[[COORD_INNER]] : (!fir.ref<!fir.type<_QMpolymorphic_testTp1{a:i32,b:i32}>>) -> !fir.class<!fir.type<_QMpolymorphic_testTp1{a:i32,b:i32}>>
-! CHECK:   %{{.*}} = fir.call @_QMpolymorphic_testPlt(%17, %[[EMBOXED]]) {{.*}} : (!fir.ref<i32>, !fir.class<!fir.type<_QMpolymorphic_testTp1{a:i32,b:i32}>>) -> !fir.logical<4>
+! CHECK:   %{{.*}} = fir.call @_QMpolymorphic_testPlt(%{{.*}}, %[[EMBOXED]]) {{.*}} : (!fir.ref<i32>, !fir.class<!fir.type<_QMpolymorphic_testTp1{a:i32,b:i32}>>) -> !fir.logical<4>
 ! CHECK:  }

flang/unittests/Runtime/CUDA/Allocatable.cpp

@@ -42,7 +42,8 @@ TEST(AllocatableCUFTest, SimpleDeviceAllocatable) {
   CUDA_REPORT_IF_ERROR(cudaMalloc(&device_desc, a->SizeInBytes()));
 
   RTNAME(AllocatableAllocate)
-  (*a, /*hasStat=*/false, /*errMsg=*/nullptr, __FILE__, __LINE__);
+  (*a, /*asyncId=*/-1, /*hasStat=*/false, /*errMsg=*/nullptr, __FILE__,
+      __LINE__);
   EXPECT_TRUE(a->IsAllocated());
   RTNAME(CUFDescriptorSync)(device_desc, a.get(), __FILE__, __LINE__);
   cudaDeviceSynchronize();

flang/unittests/Runtime/CUDA/AllocatorCUF.cpp

@@ -35,7 +35,8 @@ TEST(AllocatableCUFTest, SimpleDeviceAllocate) {
   EXPECT_FALSE(a->HasAddendum());
   RTNAME(AllocatableSetBounds)(*a, 0, 1, 10);
   RTNAME(AllocatableAllocate)
-  (*a, /*hasStat=*/false, /*errMsg=*/nullptr, __FILE__, __LINE__);
+  (*a, /*asyncId=*/-1, /*hasStat=*/false, /*errMsg=*/nullptr, __FILE__,
+      __LINE__);
   EXPECT_TRUE(a->IsAllocated());
   RTNAME(AllocatableDeallocate)
   (*a, /*hasStat=*/false, /*errMsg=*/nullptr, __FILE__, __LINE__);
@@ -53,7 +54,8 @@ TEST(AllocatableCUFTest, SimplePinnedAllocate) {
   EXPECT_FALSE(a->HasAddendum());
   RTNAME(AllocatableSetBounds)(*a, 0, 1, 10);
   RTNAME(AllocatableAllocate)
-  (*a, /*hasStat=*/false, /*errMsg=*/nullptr, __FILE__, __LINE__);
+  (*a, /*asyncId=*/-1, /*hasStat=*/false, /*errMsg=*/nullptr, __FILE__,
+      __LINE__);
   EXPECT_TRUE(a->IsAllocated());
   RTNAME(AllocatableDeallocate)
   (*a, /*hasStat=*/false, /*errMsg=*/nullptr, __FILE__, __LINE__);

flang/unittests/Runtime/CUDA/Memory.cpp

@@ -51,7 +51,8 @@ TEST(MemoryCUFTest, CUFDataTransferDescDesc) {
   EXPECT_EQ((int)kDeviceAllocatorPos, dev->GetAllocIdx());
   RTNAME(AllocatableSetBounds)(*dev, 0, 1, 10);
   RTNAME(AllocatableAllocate)
-  (*dev, /*hasStat=*/false, /*errMsg=*/nullptr, __FILE__, __LINE__);
+  (*dev, /*asyncId=*/-1, /*hasStat=*/false, /*errMsg=*/nullptr, __FILE__,
+      __LINE__);
   EXPECT_TRUE(dev->IsAllocated());
 
   // Create temp array to transfer to device.