[Clang][OpenMP] Find the type `omp_allocator_handle_t` from identifier table

[tianshilei1992]

Summary

In Clang, in order to determine the type of omp_allocator_handle_t, Clang
checks the type of those predefined allocators. The first one it checks is
omp_null_allocator. If the language is C, and the system is 64-bit, what Clang
gets is a int, instead of an enum of size 8, given the fact how we define
omp_allocator_handle_t in omp.h. If the allocator is captured by a region,
let's say a parallel region, the allocator will be privatized. Because Clang deems
omp_allocator_handle_t as an int, it will first cast the value returned by
the runtime library (for libomp it is a void *) to int, and then in the
outlined function, it casts back to omp_allocator_handle_t. This two casts
completely shaves the first 32-bit of the pointer value returned from libomp,
and when the private "new" pointer is fed to another runtime function
__kmpc_allocate(), it causes segment fault. That is the root cause of PR54082.
I have no idea why -fno-pic could hide this bug.

In this patch, we detect omp_allocator_handle_t using roughly the same method
as omp_event_handle_t, by looking it up into the identifier table.

Fix #54082.

[ABataev]

clang/test/OpenMP/target_uses_allocators.c

108

Would be good to have a test that shows the difference.

[tianshilei1992]

rebase and fix comments

[ABataev]

clang/test/OpenMP/target_uses_allocators.c

108

A unit test with the difference? This test does not show anything.

[tianshilei1992]

clang/test/OpenMP/target_uses_allocators.c

108

I added a front end test above. It shows the key difference when call __kmpc_alloc.

[tianshilei1992]

use compiler builtin for types

[ABataev]

LG

[mstorsjo]

This change broke the parallel/bug54082.c testcase in the OpenMP runtime test set, when running on Windows (both mingw and MSVC configurations, and happening both in i386 and x86_64 builds), see e.g. https://github.com/mstorsjo/llvm-mingw/actions/runs/4011290068/jobs/6891673869. It's failing with error code 0xC00000FD, which means STATUS_STACK_OVERFLOW. Is this change missing some aspect that differs between the unix and Windows ABIs?

So please don't backport this to the release branch right now. And I'd appreciate if you would consider to revert it if there's no immediately obvious fix in sight.

[tianshilei1992]

In D142297#4082242, @mstorsjo wrote:

This change broke the parallel/bug54082.c testcase in the OpenMP runtime test set, when running on Windows (both mingw and MSVC configurations, and happening both in i386 and x86_64 builds), see e.g. https://github.com/mstorsjo/llvm-mingw/actions/runs/4011290068/jobs/6891673869. It's failing with error code 0xC00000FD, which means STATUS_STACK_OVERFLOW. Is this change missing some aspect that differs between the unix and Windows ABIs?

So please don't backport this to the release branch right now. And I'd appreciate if you would consider to revert it if there's no immediately obvious fix in sight.

Thanks for the information. I'll install a Windows VM and try to fix it and get it back ported to the release branch as this patch has already been included.

clang/lib/Sema/SemaOpenMP.cpp

 getAllocatorKind(Sema &S, DSAStackTy *Stack, Expr *Allocator) {
   if (!Allocator)
     return OMPAllocateDeclAttr::OMPNullMemAlloc;
   if (Allocator->isTypeDependent() || Allocator->isValueDependent() ||
       Allocator->isInstantiationDependent() ||
       Allocator->containsUnexpandedParameterPack())
     return OMPAllocateDeclAttr::OMPUserDefinedMemAlloc;
   auto AllocatorKindRes = OMPAllocateDeclAttr::OMPUserDefinedMemAlloc;
+  llvm::FoldingSetNodeID AEId;
   const Expr *AE = Allocator->IgnoreParenImpCasts();
+  AE->IgnoreImpCasts()->Profile(AEId, S.getASTContext(), /*Canonical=*/true);
   for (int I = 0; I < OMPAllocateDeclAttr::OMPUserDefinedMemAlloc; ++I) {
     auto AllocatorKind = static_cast<OMPAllocateDeclAttr::AllocatorTypeTy>(I);
     const Expr *DefAllocator = Stack->getAllocator(AllocatorKind);
-    llvm::FoldingSetNodeID AEId, DAEId;
-    AE->Profile(AEId, S.getASTContext(), /*Canonical=*/true);
-    DefAllocator->Profile(DAEId, S.getASTContext(), /*Canonical=*/true);
+    llvm::FoldingSetNodeID DAEId;
+    DefAllocator->IgnoreImpCasts()->Profile(DAEId, S.getASTContext(),
+                                            /*Canonical=*/true);
     if (AEId == DAEId) {
       AllocatorKindRes = AllocatorKind;
       break;
     }
   }
   return AllocatorKindRes;
 }
 
if (Simdlen.isInvalid())
     return nullptr;
   return new (Context)
       OMPSimdlenClause(Simdlen.get(), StartLoc, LParenLoc, EndLoc);
 }
 
 /// Tries to find omp_allocator_handle_t type.
 static bool findOMPAllocatorHandleT(Sema &S, SourceLocation Loc,
                                     DSAStackTy *Stack) {
-  QualType OMPAllocatorHandleT = Stack->getOMPAllocatorHandleT();
-  if (!OMPAllocatorHandleT.isNull())
+  if (!Stack->getOMPAllocatorHandleT().isNull())
     return true;
-  // Build the predefined allocator expressions.
+
+  // Set the allocator handle type.
+  IdentifierInfo *II = &S.PP.getIdentifierTable().get("omp_allocator_handle_t");
+  ParsedType PT = S.getTypeName(*II, Loc, S.getCurScope());
+  if (!PT.getAsOpaquePtr() || PT.get().isNull()) {
+    S.Diag(Loc, diag::err_omp_implied_type_not_found)
+        << "omp_allocator_handle_t";
+    return false;
+  }
+  QualType AllocatorHandleEnumTy = PT.get();
+  AllocatorHandleEnumTy.addConst();
+  Stack->setOMPAllocatorHandleT(AllocatorHandleEnumTy);
+
+  // Fill the predefined allocator map.
   bool ErrorFound = false;
   for (int I = 0; I < OMPAllocateDeclAttr::OMPUserDefinedMemAlloc; ++I) {
     auto AllocatorKind = static_cast<OMPAllocateDeclAttr::AllocatorTypeTy>(I);
     StringRef Allocator =
         OMPAllocateDeclAttr::ConvertAllocatorTypeTyToStr(AllocatorKind);
     DeclarationName AllocatorName = &S.getASTContext().Idents.get(Allocator);
     auto *VD = dyn_cast_or_null<ValueDecl>(
         S.LookupSingleName(S.TUScope, AllocatorName, Loc, Sema::LookupAnyName));
     if (!VD) {
       ErrorFound = true;
       break;
     }
     QualType AllocatorType =
         VD->getType().getNonLValueExprType(S.getASTContext());
     ExprResult Res = S.BuildDeclRefExpr(VD, AllocatorType, VK_LValue, Loc);
     if (!Res.isUsable()) {
       ErrorFound = true;
       break;
     }
-    if (OMPAllocatorHandleT.isNull())
-      OMPAllocatorHandleT = AllocatorType;
-    if (!S.getASTContext().hasSameType(OMPAllocatorHandleT, AllocatorType)) {
+    Res = S.PerformImplicitConversion(Res.get(), AllocatorHandleEnumTy,
+                                      Sema::AA_Initializing,
+                                      /* AllowExplicit */ true);
+    if (!Res.isUsable()) {
       ErrorFound = true;
       break;
     }
     Stack->setAllocator(AllocatorKind, Res.get());
   }
   if (ErrorFound) {
     S.Diag(Loc, diag::err_omp_implied_type_not_found)
         << "omp_allocator_handle_t";
     return false;
   }
-  OMPAllocatorHandleT.addConst();
-  Stack->setOMPAllocatorHandleT(OMPAllocatorHandleT);
+
   return true;
 }
 
 OMPClause *Sema::ActOnOpenMPAllocatorClause(Expr *A, SourceLocation StartLoc,
                                             SourceLocation LParenLoc,
                                             SourceLocation EndLoc) {
   // OpenMP [2.11.3, allocate Directive, Description]
   // allocator is an expression of omp_allocator_handle_t type.
for (const UsesAllocatorsData &D : Data) {
     if (D.Allocator->isTypeDependent()) {
       AllocatorExpr = D.Allocator;
     } else {
       // Traits were specified - need to assign new allocator to the specified
       // allocator, so it must be an lvalue.
       AllocatorExpr = D.Allocator->IgnoreParenImpCasts();
       auto *DRE = dyn_cast<DeclRefExpr>(AllocatorExpr);
       bool IsPredefinedAllocator = false;
-      if (DRE)
-        IsPredefinedAllocator = PredefinedAllocators.count(DRE->getDecl());
-      if (!DRE ||
-          !(Context.hasSameUnqualifiedType(
-                AllocatorExpr->getType(), DSAStack->getOMPAllocatorHandleT()) ||
-            Context.typesAreCompatible(AllocatorExpr->getType(),
-                                       DSAStack->getOMPAllocatorHandleT(),
-                                       /*CompareUnqualified=*/true)) ||
-          (!IsPredefinedAllocator &&
-           (AllocatorExpr->getType().isConstant(Context) ||
-            !AllocatorExpr->isLValue()))) {
+      if (DRE) {
+        OMPAllocateDeclAttr::AllocatorTypeTy AllocatorTy =
+            getAllocatorKind(*this, DSAStack, AllocatorExpr);
+        IsPredefinedAllocator =
+            AllocatorTy !=
+            OMPAllocateDeclAttr::AllocatorTypeTy::OMPUserDefinedMemAlloc;
+      }
+      QualType OMPAllocatorHandleT = DSAStack->getOMPAllocatorHandleT();
+      QualType AllocatorExprType = AllocatorExpr->getType();
+      bool IsTypeCompatible = IsPredefinedAllocator;
+      IsTypeCompatible = IsTypeCompatible ||
+                         Context.hasSameUnqualifiedType(AllocatorExprType,
+                                                        OMPAllocatorHandleT);
+      IsTypeCompatible =
+          IsTypeCompatible ||
+          Context.typesAreCompatible(AllocatorExprType, OMPAllocatorHandleT);
+      bool IsNonConstantLValue =
+          !AllocatorExprType.isConstant(Context) && AllocatorExpr->isLValue();
+      if (!DRE || !IsTypeCompatible ||
+          (!IsPredefinedAllocator && !IsNonConstantLValue)) {
         Diag(D.Allocator->getExprLoc(), diag::err_omp_var_expected)
             << "omp_allocator_handle_t" << (DRE ? 1 : 0)
             << AllocatorExpr->getType() << D.Allocator->getSourceRange();
         continue;
       }
       // OpenMP [2.12.5, target Construct]
       // Predefined allocators appearing in a uses_allocators clause cannot have
       // traits specified.

clang/test/OpenMP/bug54082.c

+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --function-signature --include-generated-funcs --prefix-filecheck-ir-name _
+// RUN: %clang_cc1 -fopenmp -O1 -x c -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK
+
+typedef enum omp_allocator_handle_t {
+  omp_null_allocator = 0,
+  omp_default_mem_alloc = 1,
+  omp_large_cap_mem_alloc = 2,
+  omp_const_mem_alloc = 3,
+  omp_high_bw_mem_alloc = 4,
+  omp_low_lat_mem_alloc = 5,
+  omp_cgroup_mem_alloc = 6,
+  omp_pteam_mem_alloc = 7,
+  omp_thread_mem_alloc = 8,
+  KMP_ALLOCATOR_MAX_HANDLE = __UINTPTR_MAX__
+} omp_allocator_handle_t;
+
+typedef enum omp_memspace_handle_t {
+  omp_default_mem_space = 0,
+  omp_large_cap_mem_space = 1,
+  omp_const_mem_space = 2,
+  omp_high_bw_mem_space = 3,
+  omp_low_lat_mem_space = 4,
+  llvm_omp_target_host_mem_space = 100,
+  llvm_omp_target_shared_mem_space = 101,
+  llvm_omp_target_device_mem_space = 102,
+  KMP_MEMSPACE_MAX_HANDLE = __UINTPTR_MAX__
+} omp_memspace_handle_t;
+
+typedef enum {
+  omp_atk_sync_hint = 1,
+  omp_atk_alignment = 2,
+  omp_atk_access = 3,
+  omp_atk_pool_size = 4,
+  omp_atk_fallback = 5,
+  omp_atk_fb_data = 6,
+  omp_atk_pinned = 7,
+  omp_atk_partition = 8
+} omp_alloctrait_key_t;
+
+typedef __UINTPTR_TYPE__ omp_uintptr_t;
+typedef __SIZE_TYPE__ size_t;
+
+typedef struct {
+  omp_alloctrait_key_t key;
+  omp_uintptr_t value;
+} omp_alloctrait_t;
+
+extern omp_allocator_handle_t
+omp_init_allocator(omp_memspace_handle_t memspace, int ntraits,
+                   const omp_alloctrait_t traits[]);
+
+#define N 1024
+
+void foo() {
+  int *x;
+
+  omp_memspace_handle_t x_memspace = omp_default_mem_space;
+  omp_alloctrait_t x_traits[1] = {omp_atk_alignment, 64};
+  omp_allocator_handle_t x_alloc = omp_init_allocator(x_memspace, 1, x_traits);
+
+#pragma omp parallel for allocate(x_alloc : x) private(x)
+  for (int i = 0; i < N; i++) {
+    (void)x;
+  }
+}
+// CHECK-LABEL: define {{[^@]+}}@foo
+// CHECK-SAME: () local_unnamed_addr #[[ATTR0:[0-9]+]] {
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[X_TRAITS:%.*]] = alloca [1 x %struct.omp_alloctrait_t], align 16
+// CHECK-NEXT:    [[X_ALLOC:%.*]] = alloca i64, align 8
+// CHECK-NEXT:    call void @llvm.lifetime.start.p0(i64 16, ptr nonnull [[X_TRAITS]]) #[[ATTR5:[0-9]+]]
+// CHECK-NEXT:    call void @llvm.memcpy.p0.p0.i64(ptr noundef nonnull align 16 dereferenceable(16) [[X_TRAITS]], ptr noundef nonnull align 16 dereferenceable(16) @__const.foo.x_traits, i64 16, i1 false)
+// CHECK-NEXT:    call void @llvm.lifetime.start.p0(i64 8, ptr nonnull [[X_ALLOC]]) #[[ATTR5]]
+// CHECK-NEXT:    [[CALL:%.*]] = call i64 @omp_init_allocator(i64 noundef 0, i32 noundef 1, ptr noundef nonnull [[X_TRAITS]]) #[[ATTR5]]
+// CHECK-NEXT:    store i64 [[CALL]], ptr [[X_ALLOC]], align 8, !tbaa [[TBAA3:![0-9]+]]
+// CHECK-NEXT:    call void (ptr, i32, ptr, ...) @__kmpc_fork_call(ptr nonnull @[[GLOB2:[0-9]+]], i32 1, ptr nonnull @.omp_outlined., ptr nonnull [[X_ALLOC]])
+// CHECK-NEXT:    call void @llvm.lifetime.end.p0(i64 8, ptr nonnull [[X_ALLOC]]) #[[ATTR5]]
+// CHECK-NEXT:    call void @llvm.lifetime.end.p0(i64 16, ptr nonnull [[X_TRAITS]]) #[[ATTR5]]
+// CHECK-NEXT:    ret void
+//
+//
+// CHECK-LABEL: define {{[^@]+}}@.omp_outlined.
+// CHECK-SAME: (ptr noalias nocapture noundef readonly [[DOTGLOBAL_TID_:%.*]], ptr noalias nocapture readnone [[DOTBOUND_TID_:%.*]], ptr nocapture noundef nonnull readonly align 8 dereferenceable(8) [[X_ALLOC:%.*]]) #[[ATTR4:[0-9]+]] {
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[DOTOMP_LB:%.*]] = alloca i32, align 4
+// CHECK-NEXT:    [[DOTOMP_UB:%.*]] = alloca i32, align 4
+// CHECK-NEXT:    [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
+// CHECK-NEXT:    [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
+// CHECK-NEXT:    call void @llvm.lifetime.start.p0(i64 4, ptr nonnull [[DOTOMP_LB]]) #[[ATTR5]]
+// CHECK-NEXT:    store i32 0, ptr [[DOTOMP_LB]], align 4, !tbaa [[TBAA6:![0-9]+]]
+// CHECK-NEXT:    call void @llvm.lifetime.start.p0(i64 4, ptr nonnull [[DOTOMP_UB]]) #[[ATTR5]]
+// CHECK-NEXT:    store i32 1023, ptr [[DOTOMP_UB]], align 4, !tbaa [[TBAA6]]
+// CHECK-NEXT:    call void @llvm.lifetime.start.p0(i64 4, ptr nonnull [[DOTOMP_STRIDE]]) #[[ATTR5]]
+// CHECK-NEXT:    store i32 1, ptr [[DOTOMP_STRIDE]], align 4, !tbaa [[TBAA6]]
+// CHECK-NEXT:    call void @llvm.lifetime.start.p0(i64 4, ptr nonnull [[DOTOMP_IS_LAST]]) #[[ATTR5]]
+// CHECK-NEXT:    store i32 0, ptr [[DOTOMP_IS_LAST]], align 4, !tbaa [[TBAA6]]
+// CHECK-NEXT:    [[TMP0:%.*]] = load i32, ptr [[DOTGLOBAL_TID_]], align 4, !tbaa [[TBAA6]]
+// CHECK-NEXT:    [[TMP1:%.*]] = load i64, ptr [[X_ALLOC]], align 8, !tbaa [[TBAA3]]
+// CHECK-NEXT:    [[CONV:%.*]] = inttoptr i64 [[TMP1]] to ptr
+// CHECK-NEXT:    [[DOTX__VOID_ADDR:%.*]] = tail call ptr @__kmpc_alloc(i32 [[TMP0]], i64 8, ptr [[CONV]])
+// CHECK-NEXT:    call void @__kmpc_for_static_init_4(ptr nonnull @[[GLOB1:[0-9]+]], i32 [[TMP0]], i32 34, ptr nonnull [[DOTOMP_IS_LAST]], ptr nonnull [[DOTOMP_LB]], ptr nonnull [[DOTOMP_UB]], ptr nonnull [[DOTOMP_STRIDE]], i32 1, i32 1)
+// CHECK-NEXT:    [[TMP2:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4, !tbaa [[TBAA6]]
+// CHECK-NEXT:    [[COND:%.*]] = call i32 @llvm.smin.i32(i32 [[TMP2]], i32 1023)
+// CHECK-NEXT:    store i32 [[COND]], ptr [[DOTOMP_UB]], align 4, !tbaa [[TBAA6]]
+// CHECK-NEXT:    call void @__kmpc_for_static_fini(ptr nonnull @[[GLOB1]], i32 [[TMP0]])
+// CHECK-NEXT:    [[TMP3:%.*]] = load i64, ptr [[X_ALLOC]], align 8, !tbaa [[TBAA3]]
+// CHECK-NEXT:    [[CONV5:%.*]] = inttoptr i64 [[TMP3]] to ptr
+// CHECK-NEXT:    call void @__kmpc_free(i32 [[TMP0]], ptr [[DOTX__VOID_ADDR]], ptr [[CONV5]])
+// CHECK-NEXT:    call void @llvm.lifetime.end.p0(i64 4, ptr nonnull [[DOTOMP_IS_LAST]]) #[[ATTR5]]
+// CHECK-NEXT:    call void @llvm.lifetime.end.p0(i64 4, ptr nonnull [[DOTOMP_STRIDE]]) #[[ATTR5]]
+// CHECK-NEXT:    call void @llvm.lifetime.end.p0(i64 4, ptr nonnull [[DOTOMP_UB]]) #[[ATTR5]]
+// CHECK-NEXT:    call void @llvm.lifetime.end.p0(i64 4, ptr nonnull [[DOTOMP_LB]]) #[[ATTR5]]
+// CHECK-NEXT:    ret void
+//

clang/test/OpenMP/target_uses_allocators.c

 // Test host codegen.
 // RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=50  -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
 // RUN: %clang_cc1 -fopenmp -fopenmp-version=50 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -verify -emit-pch -o %t %s
 // RUN: %clang_cc1 -fopenmp -fopenmp-version=50 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -include-pch %t %s -emit-llvm -o - | FileCheck %s
 
 #ifndef HEADER
 #define HEADER
 
-enum omp_allocator_handle_t {
+typedef enum omp_allocator_handle_t {
   omp_null_allocator = 0,
   omp_default_mem_alloc = 1,
   omp_large_cap_mem_alloc = 2,
   omp_const_mem_alloc = 3,
   omp_high_bw_mem_alloc = 4,
   omp_low_lat_mem_alloc = 5,
   omp_cgroup_mem_alloc = 6,
   omp_pteam_mem_alloc = 7,
   omp_thread_mem_alloc = 8,
   KMP_ALLOCATOR_MAX_HANDLE = __UINTPTR_MAX__
-};
+} omp_allocator_handle_t;
 
 // CHECK: define {{.*}}[[FIE:@.+]]()
 void fie(void) {
   int x;
   #pragma omp target uses_allocators(omp_null_allocator) allocate(omp_null_allocator: x) firstprivate(x)
   {}
   #pragma omp target uses_allocators(omp_default_mem_alloc) allocate(omp_default_mem_alloc: x) firstprivate(x)
   {}

 // CHECK-NEXT: store i32 %[[#R1]], ptr %.x..void.addr, align 4
 // CHECK-NEXT: call void @__kmpc_free(i32 %[[#R0]], ptr %.x..void.addr, ptr inttoptr (i64 7 to ptr))
 
 // CHECK: %[[#R0:]] = call i32 @__kmpc_global_thread_num(ptr @1)
 // CHECK-NEXT: store i64 %x, ptr %x.addr, align 8
 // CHECK-NEXT: %.x..void.addr = call ptr @__kmpc_alloc(i32 %[[#R0]], i64 4, ptr inttoptr (i64 8 to ptr))
 // CHECK-NEXT: %[[#R1:]] = load i32, ptr %x.addr, align 4
 // CHECK-NEXT: store i32 %[[#R1]], ptr %.x..void.addr, align 4
-// CHECK-NEXT: call void @__kmpc_free(i32 %[[#R0]], ptr %.x..void.addr, ptr inttoptr (i64 8 to ptr))
+// CHECK-NEXT: call void @__kmpc_free(i32 %[[#R0]], ptr %.x..void.addr, ptr inttoptr (i64 8 to ptr))
 

openmp/runtime/test/parallel/bug54082.c

+// This test is adapted from test_parallel_for_allocate.c in SOLLVE V&V.
+// https://github.com/SOLLVE/sollve_vv/blob/master/tests/5.0/parallel_for/test_parallel_for_allocate.c
+// RUN: %libomp-compile-and-run
+#include <omp.h>
+
+#include <assert.h>
+#include <stdlib.h>
+
+#define N 1024
+
+int main(int argc, char *argv[]) {
+  int errors = 0;
+  int *x;
+  int result[N][N];
+  int successful_alloc = 0;
+
+  omp_memspace_handle_t x_memspace = omp_default_mem_space;
+  omp_alloctrait_t x_traits[1] = {omp_atk_alignment, 64};
+  omp_allocator_handle_t x_alloc = omp_init_allocator(x_memspace, 1, x_traits);
+
+  for (int i = 0; i < N; i++) {
+    for (int j = 0; j < N; j++) {
+      result[i][j] = -1;
+    }
+  }
+
+#pragma omp parallel for allocate(x_alloc: x) private(x) shared(result)
+  for (int i = 0; i < N; i++) {
+    x = (int *)malloc(N * sizeof(int));
+    if (x != NULL) {
+#pragma omp simd simdlen(16) aligned(x : 64)
+      for (int j = 0; j < N; j++) {
+        x[j] = j * i;
+      }
+      for (int j = 0; j < N; j++) {
+        result[i][j] = x[j];
+      }
+      free(x);
+      successful_alloc++;
+    }
+  }
+
+  errors += successful_alloc < 1;
+
+  for (int i = 0; i < N; i++) {
+    for (int j = 0; j < N; j++) {
+      errors += result[i][j] != i * j;
+    }
+  }
+
+  omp_destroy_allocator(x_alloc);
+
+  return errors;
+}