Merge pull request #2 from siemens/master

Pull master from Siemens/embb

CHANGELOG.md

 Embedded Multicore Building Blocks (EMB²)
 =========================================
 
+Version 0.3.1
+-------------
+
+### Features:
+- None
+
+### Changes and improvements:
+- Removed one function argument from algorithms::Invoke
+- Added "explicit" specifier to base type constructor of Atomic<BaseType*>
+- Added "const" qualifier to dereference operator and member access operator of AtomicPointer<>
+- Changed AtomicBase<>::CompareAndSwap to atomically return expected value
+- Replaced constant in dataflow_cpp_test_simple.cc with corresponding macro
+- Added initialization of atomic variable in hazard_pointer_test.cc to avoid warning with GCC 5.1
+- Changed initial value of allocated_object_from_different_thread
+- Added tests for ID Pool and check for memory leaks
+- Updated unit test for the UniqueLock::Swap
+
+### Bug fixes:
+- Fixed implementation of ID pool (provided fewer elements than specified by capacity)
+- Fixed unsigned overflow bug in timed wait function of condition variables
+- Fixed implementation of UniqueLock::Swap
+
+### Build system:
+- Improved CMake output for automatic initialization option
+- Fixed cpplint and unsigned/signed warnings
+
+### Documentation:
+- Fixed documentation of UniqueLock class
+- Updated README file
+
+
 Version 0.3.0
 -------------
 

CMakeLists.txt

@@ -28,7 +28,7 @@ cmake_minimum_required (VERSION 2.8.9)
 # Version number
 set (EMBB_BASE_VERSION_MAJOR 0)
 set (EMBB_BASE_VERSION_MINOR 3)
-set (EMBB_BASE_VERSION_PATCH 0)
+set (EMBB_BASE_VERSION_PATCH 1)
 
 # Fix compilation for CMake versions >= 3.1
 #
@@ -100,6 +100,13 @@ else()
 endif()
 message("   (set with command line option -DWARNINGS_ARE_ERRORS=ON/OFF)")
 
+if (USE_AUTOMATIC_INITIALIZATION STREQUAL ON)
+  message("-- MTAPI/Tasks automatic initialization enabled (default)")
+else()
+  message("-- MTAPI/Tasks automatic initialization disabled")
+endif()
+message("   (set with command line option -DUSE_AUTOMATIC_INITIALIZATION=ON/OFF)")
+
 include(CMakeCommon/SetCompilerFlags.cmake)
 SetGNUCompilerFlags(compiler_libs compiler_flags)
 SetVisualStudioCompilerFlags(compiler_libs compiler_flags)

README.md

@@ -323,6 +323,8 @@ Known Bugs and Limitations
   is bounded by a predefined but modifiable constant (see functions
   embb_thread_get_max_count() / embb_thread_set_max_count() and class
   embb::base::Thread).
+- While MTAPI fully supports heterogeneous systems, the algorithms and
+  dataflow components are currently limited to homogeneous systems.
 
 
 Development and Contribution

algorithms_cpp/include/embb/algorithms/invoke.h

@@ -49,33 +49,37 @@ typedef embb::base::Function<void> InvokeFunctionType;
 #ifdef DOXYGEN
 
 /**
- * Spawns one to ten function objects at once and runs them in parallel.
+ * Spawns two to ten function objects at once and runs them in parallel.
  *
  * Blocks until all of them are done.
  *
  * \ingroup CPP_ALGORITHMS_INVOKE
  */
-template<typename Function1, ...>
+template<typename Function1, typename Function2, ...>
 void Invoke(
   Function1 func1,
   /**< [in] First function object to invoke */
+  Function2 func2,
+  /**< [in] Second function object to invoke */
   ...);
 
 /**
-* Spawns one to ten function objects at once and runs them in parallel using the
+* Spawns two to ten function objects at once and runs them in parallel using the
 * given embb::mtapi::ExecutionPolicy.
 *
 * Blocks until all of them are done.
 *
 * \ingroup CPP_ALGORITHMS_INVOKE
 */
-template<typename Function1, ...>
+template<typename Function1, typename Function2, ...>
 void Invoke(
   Function1 func1,
   /**< [in] Function object to invoke */
+  Function2 func2,
+  /**< [in] Second function object to invoke */
   ...,
-  const embb::mtapi::ExecutionPolicy & policy
-  /**< [in] embb::mtapi::ExecutionPolicy to use */
+  const embb::tasks::ExecutionPolicy & policy
+  /**< [in] embb::tasks::ExecutionPolicy to use */
   );
 
 #else // DOXYGEN
@@ -118,13 +122,6 @@ class TaskWrapper {
 };
 } // namespace internal
 
-template<typename Function1>
-void Invoke(
-  Function1 func1,
-  const embb::tasks::ExecutionPolicy& policy) {
-  internal::TaskWrapper<Function1> wrap1(func1, policy);
-}
-
 template<typename Function1, typename Function2>
 void Invoke(
   Function1 func1,
@@ -290,12 +287,6 @@ template<typename Function1, typename Function2, typename Function3,
   internal::TaskWrapper<Function10> wrap10(func10, policy);
 }
 
-template<typename Function1>
-void Invoke(
-  Function1 func1) {
-  Invoke(func1, embb::tasks::ExecutionPolicy());
-}
-
 template<typename Function1, typename Function2>
 void Invoke(
   Function1 func1,

algorithms_cpp/test/invoke_test.cc

@@ -44,7 +44,6 @@ static void Invocable10() {}
 
 void InvokeTest::Test() {
   using embb::algorithms::Invoke;
-  Invoke(&Invocable1);
   Invoke(&Invocable1, &Invocable2);
   Invoke(&Invocable1, &Invocable2, &Invocable3);
   Invoke(&Invocable1, &Invocable2, &Invocable3, &Invocable4);
@@ -61,4 +60,24 @@ void InvokeTest::Test() {
          &Invocable6, &Invocable7, &Invocable8, &Invocable9);
   Invoke(&Invocable1, &Invocable2, &Invocable3, &Invocable4, &Invocable5,
          &Invocable6, &Invocable7, &Invocable8, &Invocable9, &Invocable10);
+
+  embb::tasks::ExecutionPolicy policy;
+  Invoke(&Invocable1, &Invocable2, policy);
+  Invoke(&Invocable1, &Invocable2, &Invocable3, policy);
+  Invoke(&Invocable1, &Invocable2, &Invocable3, &Invocable4, policy);
+  Invoke(&Invocable1, &Invocable2, &Invocable3, &Invocable4, &Invocable5,
+         policy);
+  Invoke(&Invocable1, &Invocable2, &Invocable3, &Invocable4, &Invocable5,
+         &Invocable6, policy);
+  Invoke(&Invocable1, &Invocable2, &Invocable3, &Invocable4, &Invocable5,
+         &Invocable6, &Invocable7, policy);
+  Invoke(&Invocable1, &Invocable2, &Invocable3, &Invocable4, &Invocable5,
+         &Invocable6, &Invocable7, &Invocable8, policy);
+  Invoke(&Invocable1, &Invocable2, &Invocable3, &Invocable4, &Invocable5,
+         &Invocable6, &Invocable7, &Invocable8, &Invocable9, policy);
+  Invoke(&Invocable1, &Invocable2, &Invocable3, &Invocable4, &Invocable5,
+         &Invocable6, &Invocable7, &Invocable8, &Invocable9, policy);
+  Invoke(&Invocable1, &Invocable2, &Invocable3, &Invocable4, &Invocable5,
+         &Invocable6, &Invocable7, &Invocable8, &Invocable9, &Invocable10,
+         policy);
 }

base_c/src/condition_variable.c

@@ -83,8 +83,8 @@ int embb_condition_wait_until(embb_condition_t* condition_var,
   embb_time_t now;
   embb_time_now(&now);
   /* Check if absolute timepoint (in milliseconds) still is in the future */
-  if (time->seconds * 1000 + time->nanoseconds / 1000000
-      - now.seconds * 1000 - now.nanoseconds / 1000000 > 0) {
+  if ((time->seconds * 1000 + time->nanoseconds / 1000000)
+      > (now.seconds * 1000 + now.nanoseconds / 1000000)) {
     /* Convert to (unsigned type) milliseconds and round up */
     DWORD time_diff = (DWORD) (
         time->seconds * 1000 + time->nanoseconds / 1000000

base_c/test/condition_var_test.cc

@@ -105,7 +105,7 @@ void ConditionVarTest::TestTimedWaitTimeouts() {
   embb_time_t time;
   embb_duration_t duration = EMBB_DURATION_INIT;
 
-  // Wait for now tests already passed time point
+  // Wait for "now" tests already passed time point
   embb_time_now(&time);
   embb_mutex_lock(&mutex);
   int status = embb_condition_wait_until(&cond, &mutex, &time);

base_c/test/time_test.cc

@@ -36,6 +36,9 @@ namespace test {
 
 TimeTest::TimeTest() {
   CreateUnit("Time in duration").Add(&TimeTest::TestTimeInDuration, this);
+  CreateUnit("Monotonicity").Add(
+      &TimeTest::TestMonotonicity, this,
+      1, partest::TestSuite::GetDefaultNumIterations() * 10);
 }
 
 void TimeTest::TestTimeInDuration() {
@@ -48,6 +51,20 @@ void TimeTest::TestTimeInDuration() {
   PT_EXPECT_EQ(status, EMBB_SUCCESS);
 }
 
+void TimeTest::TestMonotonicity() {
+  embb_time_t first;
+  embb_time_t second;
+  int status1 = embb_time_in(&first, embb_duration_zero());
+  int status2 = embb_time_in(&second, embb_duration_zero());
+  PT_EXPECT_EQ(status1, EMBB_SUCCESS);
+  PT_EXPECT_EQ(status2, EMBB_SUCCESS);
+  unsigned long long first_abs = first.seconds * 1000 +
+                                 first.nanoseconds / 1000000;
+  unsigned long long second_abs = second.seconds * 1000 +
+                                  second.nanoseconds / 1000000;
+  PT_EXPECT_GE(second_abs, first_abs);
+}
+
 } // namespace test
 } // namespace base
 } // namespace embb

base_c/test/time_test.h

@@ -42,9 +42,14 @@ class TimeTest : public partest::TestCase {
 
  private:
   /**
-   * Tests time in duration method.
+   * Tests time-in-duration method.
    */
   void TestTimeInDuration();
+
+  /**
+   * Tests that succeedingly taken times are monotonously increasing.
+   */
+  void TestMonotonicity();
 };
 
 } // namespace test

base_cpp/include/embb/base/atomic.h

@@ -478,7 +478,7 @@ class Atomic<BaseType*> : public embb::base::internal::atomic::
  public:
   Atomic() : embb::base::internal::atomic::
     AtomicPointer<BaseType, ptrdiff_t, sizeof(BaseType*)>() {}
-  Atomic(BaseType* p) : embb::base::internal::atomic::
+  explicit Atomic(BaseType* p) : embb::base::internal::atomic::
     AtomicPointer<BaseType, ptrdiff_t, sizeof(BaseType*)>(p) {}
 
   BaseType* operator=(BaseType* p) {

base_cpp/include/embb/base/internal/atomic/atomic_base.h

@@ -177,8 +177,7 @@ CompareAndSwap(BaseType& expected, BaseType desired) {
     compare_and_swap(&AtomicValue, &native_expected, native_desired)) !=0
     ? true : false;
 
-  if (!return_val)
-    expected = Load();
+  memcpy(&expected, &native_expected, sizeof(expected));
 
   return return_val;
 }

base_cpp/include/embb/base/internal/atomic/atomic_pointer.h

@@ -65,8 +65,8 @@ class AtomicPointer : public AtomicArithmetic<BaseType*, DifferenceType, S> {
   bool IsPointer() const;
 
   // The methods below are documented in atomic.h
-  BaseType* operator->();
-  BaseType& operator*();
+  BaseType* operator->() const;
+  BaseType& operator*() const;
 };
 
 template<typename BaseType, typename DifferenceType, size_t S>
@@ -93,13 +93,13 @@ inline bool AtomicPointer<BaseType, DifferenceType, S>::
 
 template<typename BaseType, typename DifferenceType, size_t S>
 inline BaseType* AtomicPointer<BaseType, DifferenceType, S>::
-  operator->() {
+  operator->() const {
   return this->Load();
 }
 
 template<typename BaseType, typename DifferenceType, size_t S>
 inline BaseType& AtomicPointer<BaseType, DifferenceType, S>::
-  operator*() {
+  operator*() const {
   return *(this->Load());
 }
 

base_cpp/include/embb/base/internal/mutex-inl.h

@@ -28,6 +28,7 @@
 #define EMBB_BASE_INTERNAL_MUTEX_INL_H_
 
 #include <cassert>
+#include <algorithm>
 
 namespace embb {
 namespace base {
@@ -95,8 +96,8 @@ void UniqueLock<Mutex>::Unlock() {
 
 template<typename Mutex>
 void UniqueLock<Mutex>::Swap(UniqueLock<Mutex>& other) {
-  locked_ = other.locked_;
-  mutex_ = other.Release();
+  std::swap(mutex_, other.mutex_);
+  std::swap(locked_, other.locked_);
 }
 
 template<typename Mutex>

base_cpp/include/embb/base/mutex.h

@@ -439,11 +439,11 @@ class UniqueLock {
   void Unlock();
 
   /**
-   * Transfers ownership of a mutex to this lock.
+   * Exchanges ownership of the wrapped mutex with another lock.
    */
   void Swap(
     UniqueLock<Mutex>& other
-    /**< [IN/OUT] Lock from which ownership shall be transferred */
+    /**< [IN/OUT] The lock to exchange ownership with */
     );
 
   /**

base_cpp/test/mutex_test.cc

@@ -191,13 +191,21 @@ void MutexTest::TestUniqueLock() {
   }
 
   { // Test lock swapping
-    UniqueLock<> lock1;
-    UniqueLock<> lock2(mutex_);
-    PT_EXPECT_EQ(lock1.OwnsLock(), false);
-    PT_EXPECT_EQ(lock2.OwnsLock(), true);
-    lock1.Swap(lock2);
+    UniqueLock<> lock1(mutex_);
     PT_EXPECT_EQ(lock1.OwnsLock(), true);
-    PT_EXPECT_EQ(lock2.OwnsLock(), false);
+
+    {
+      UniqueLock<> lock2;
+      PT_EXPECT_EQ(lock2.OwnsLock(), false);
+
+      lock1.Swap(lock2);
+      PT_EXPECT_EQ(lock1.OwnsLock(), false);
+      PT_EXPECT_EQ(lock2.OwnsLock(), true);
+    }
+
+    // At this point, "lock2" was destroyed and "mutex_" must be unlocked.
+    UniqueLock<> lock3(mutex_, embb::base::try_lock);
+    PT_EXPECT_EQ(lock3.OwnsLock(), true);
   }
 }
 

containers_cpp/test/hazard_pointer_test.cc

@@ -97,7 +97,7 @@ void HazardPointerTest::HazardPointerTest1_ThreadMethod() {
 
     PT_ASSERT(success == true);
 
-    embb::base::Atomic<int>* allocated_object_from_different_thread;
+    embb::base::Atomic<int>* allocated_object_from_different_thread(0);
 
     int diff_count = 0;
 

dataflow_cpp/test/dataflow_cpp_test_simple.cc

@@ -39,7 +39,7 @@
 #define NUM_SLICES 8
 #define TEST_COUNT 12
 
-typedef embb::dataflow::Network<8> MyNetwork;
+typedef embb::dataflow::Network<NUM_SLICES> MyNetwork;
 typedef MyNetwork::ConstantSource< int > MyConstantSource;
 typedef MyNetwork::Source< int > MySource;
 typedef MyNetwork::SerialProcess< MyNetwork::Inputs<int>::Type,
@@ -156,9 +156,7 @@ void SimpleTest::TestBasic() {
     core_set,
     1024, // max tasks (default: 1024)
     128,  // max groups (default: 128)
-    // Currently needs to be initialized
-    // with (max_queues + 1), see defect embb449
-    num_cores + 1, // max queues (default: 16)
+    num_cores, // max queues (default: 16)
     1024, // queue capacity (default: 1024)
     4);   // num priorities (default: 4)
 

mtapi_c/src/embb_mtapi_id_pool_t.c

@@ -71,7 +71,7 @@ mtapi_uint_t embb_mtapi_id_pool_allocate(embb_mtapi_id_pool_t * that) {
       /* acquire position to fetch id from */
       mtapi_uint_t id_position = that->get_id_position;
       that->get_id_position++;
-      if (that->capacity <= that->get_id_position) {
+      if (that->capacity < that->get_id_position) {
         that->get_id_position = 0;
       }
 
@@ -97,7 +97,7 @@ void embb_mtapi_id_pool_deallocate(
       /* acquire position to put id to */
       mtapi_uint_t id_position = that->put_id_position;
       that->put_id_position++;
-      if (that->capacity <= that->put_id_position) {
+      if (that->capacity < that->put_id_position) {
         that->put_id_position = 0;
       }
 

mtapi_c/test/embb_mtapi_test_id_pool.cc

+/*
+ * Copyright (c) 2014-2015, Siemens AG. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice,
+ * this list of conditions and the following disclaimer.
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ * this list of conditions and the following disclaimer in the documentation
+ * and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <embb_mtapi_test_id_pool.h>
+#include <vector>
+
+IdPoolTest::IdPoolTest() {
+  CreateUnit("mtapi id pool test single threaded").
+    Add(&IdPoolTest::TestBasic, this, 1, 1000).
+    Pre(&IdPoolTest::TestBasicPre, this).
+    Post(&IdPoolTest::TestBasicPost, this);
+
+  CreateUnit("mtapi id pool test concurrent").
+    Add(&IdPoolTest::TestParallel, this, concurrent_accessors_id_pool_2
+    , 20).
+    Post(&IdPoolTest::TestParallelPost, this).
+    Pre(&IdPoolTest::TestParallelPre, this);
+}
+
+void IdPoolTest::TestParallel() {
+  // allocate ID_ELEMENTS_PER_ACCESSOR elements. Each test thread is
+  // guaranteed to be able to allocate this amount of elements.
+  TestAllocateDeallocateNElementsFromPool(id_pool_parallel,
+    id_elements_per_accessor);
+}
+
+void IdPoolTest::TestParallelPre() {
+  // create second id pool with CONCURRENT_ACCESSORS_ID_POOL_2*
+  // ID_ELEMENTS_PER_ACCESSOR elements
+  embb_mtapi_id_pool_initialize(&id_pool_parallel,
+    concurrent_accessors_id_pool_2*id_elements_per_accessor);
+}
+
+void IdPoolTest::TestParallelPost() {
+  // after the parallel tests, try to again allocate and deallocate all
+  // elements sequentially.
+  TestAllocateDeallocateNElementsFromPool(id_pool_parallel,
+    concurrent_accessors_id_pool_2*id_elements_per_accessor, true);
+
+  // finalize pool
+  embb_mtapi_id_pool_finalize(&id_pool_parallel);
+}
+
+void IdPoolTest::TestBasic() {
+  TestAllocateDeallocateNElementsFromPool(id_pool, id_pool_size_1, true);
+}
+
+void IdPoolTest::TestBasicPre() {
+  // create id pool with ID_POOL_SIZE_1 elements
+  embb_mtapi_id_pool_initialize(&id_pool, id_pool_size_1);
+}
+
+void IdPoolTest::TestBasicPost() {
+  // finalize pool
+  embb_mtapi_id_pool_finalize(&id_pool);
+}
+
+void IdPoolTest::TestAllocateDeallocateNElementsFromPool(
+  embb_mtapi_id_pool_t &pool,
+  int count_elements,
+  bool empty_check) {
+  std::vector<unsigned int> allocated;
+
+  for (int i = 0; i != count_elements; ++i) {
+    allocated.push_back(embb_mtapi_id_pool_allocate(&pool));
+  }
+
+  // the allocated elements should be disjunctive, and never invalid element
+  for (unsigned int x = 0; x != allocated.size(); ++x) {
+    PT_ASSERT(allocated[x] != EMBB_MTAPI_IDPOOL_INVALID_ID);
+    for (unsigned int y = 0; y != allocated.size(); ++y) {
+      if (x == y) {
+        continue;
+      }
+      PT_ASSERT(allocated[x] != allocated[y]);
+    }
+  }
+
+  // now the id pool should be empty... try ten times to get an id,
+  // we should always get the invalid element
+  if (empty_check) {
+    for (int i = 0; i != 10; ++i) {
+      PT_ASSERT_EQ(embb_mtapi_id_pool_allocate(&pool),
+        static_cast<unsigned int>(EMBB_MTAPI_IDPOOL_INVALID_ID)
+        )
+    }
+  }
+
+  // now return allocated elements in a shuffled manner.
+  ::std::random_shuffle(allocated.begin(), allocated.end());
+
+  for (int i = 0; i != count_elements; ++i) {
+    embb_mtapi_id_pool_deallocate(&pool,
+      allocated[static_cast<unsigned int>(i)]);
+  }
+}
+

mtapi_c/test/embb_mtapi_test_id_pool.h

+/*
+ * Copyright (c) 2014-2015, Siemens AG. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice,
+ * this list of conditions and the following disclaimer.
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ * this list of conditions and the following disclaimer in the documentation
+ * and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef MTAPI_C_TEST_EMBB_MTAPI_TEST_ID_POOL_H_
+#define MTAPI_C_TEST_EMBB_MTAPI_TEST_ID_POOL_H_
+
+#include <partest/partest.h>
+#include <embb_mtapi_id_pool_t.h>
+
+// for shuffling a vector
+#include <algorithm>
+
+class IdPoolTest : public partest::TestCase {
+ public:
+  embb_mtapi_id_pool_t id_pool;
+  embb_mtapi_id_pool_t id_pool_parallel;
+
+  IdPoolTest();
+
+ private:
+  static const unsigned int id_pool_size_1 = 100;
+  static const unsigned int concurrent_accessors_id_pool_2 = 10;
+  static const unsigned int id_elements_per_accessor = 10;
+
+  /**  
+   * We create a pool of size number_accessors*elements_per_accessor, so
+   * at each time we can guarantee each thread to be able to allocate 
+   * elements_per_accessor elements.
+   * We create number_accessor threads, where each thread iteratively 
+   * allocates and frees elements_per_accessor elements, which in each case
+   * has to be successful. Additionally, the sanity checks from the basic tests
+   * are repeated. The TestParallelPost function also repeats all
+   * sequential tests.
+   */
+  void TestParallel();
+  void TestParallelPre();
+  void TestParallelPost();
+
+  /**
+   * Create a pool of size N. We repeatedly allocate and free N elements, check
+   * if the pool always returns disjunctive ids and check that the pool never
+   * returns the invalid element, if the pool is not empty. Check that the 
+   * invalid element is returned if the pool is empty.
+   */
+  void TestBasic();
+  void TestBasicPre();
+  void TestBasicPost();
+
+  static void TestAllocateDeallocateNElementsFromPool(
+    embb_mtapi_id_pool_t &pool,
+    int count_elements,
+    bool empty_check = false);
+};
+
+#endif // MTAPI_C_TEST_EMBB_MTAPI_TEST_ID_POOL_H_

mtapi_c/test/main.cc

@@ -37,6 +37,9 @@
 #include <embb_mtapi_test_group.h>
 #include <embb_mtapi_test_queue.h>
 #include <embb_mtapi_test_error.h>
+#include <embb_mtapi_test_id_pool.h>
+
+#include <embb/base/c/memory_allocation.h>
 
 PT_MAIN("MTAPI C") {
   embb_log_set_log_level(EMBB_LOG_LEVEL_NONE);
@@ -48,4 +51,7 @@ PT_MAIN("MTAPI C") {
   PT_RUN(InitFinalizeTest);
   PT_RUN(GroupTest);
   PT_RUN(QueueTest);
+  PT_RUN(IdPoolTest);
+
+  PT_EXPECT(embb_get_bytes_allocated() == 0);
 }

mtapi_cpp/CMakeLists.txt

@@ -5,14 +5,10 @@ file(GLOB_RECURSE EMBB_MTAPI_CPP_HEADERS "include/*.h")
 file(GLOB_RECURSE EMBB_MTAPI_CPP_TEST_SOURCES "test/*.cc" "test/*.h")
 
 if (USE_AUTOMATIC_INITIALIZATION STREQUAL ON)
-  message("-- Automatic initialization enabled (default)")
   set(MTAPI_CPP_AUTOMATIC_INITIALIZE 1)
 else()
   set(MTAPI_CPP_AUTOMATIC_INITIALIZE 0)
-  message("-- Automatic initialization disabled")
 endif()
-message("   (set with command line option -DUSE_AUTOMATIC_INITIALIZATION=ON/OFF)")
-
 
 # Execute the GroupSources macro
 include(${CMAKE_SOURCE_DIR}/CMakeCommon/GroupSourcesMSVC.cmake)

tasks_cpp/CMakeLists.txt

@@ -5,13 +5,10 @@ file(GLOB_RECURSE EMBB_TASKS_CPP_HEADERS "include/*.h")
 file(GLOB_RECURSE EMBB_TASKS_CPP_TEST_SOURCES "test/*.cc" "test/*.h")
 
 if (USE_AUTOMATIC_INITIALIZATION STREQUAL ON)
-  message("-- Automatic initialization enabled (default)")
   set(TASKS_CPP_AUTOMATIC_INITIALIZE 1)
 else()
   set(TASKS_CPP_AUTOMATIC_INITIALIZE 0)
-  message("-- Automatic initialization disabled")
 endif()
-message("   (set with command line option -DUSE_AUTOMATIC_INITIALIZATION=ON/OFF)")
 
 configure_file("include/embb/tasks/internal/cmake_config.h.in"
                "include/embb/tasks/internal/cmake_config.h")