mpi/src/starpu_mpi.c: when dealing with an application request which

is connected to an internal request (data received early), we need to
wait for the internal request to be ready before processing it.

The patch seems to lead to some deadlocks when running examples/stencil.
This needs to be fixed.

mpi/src/starpu_mpi.c

@@ -69,6 +69,9 @@ LIST_TYPE(_starpu_mpi_copy_handle,
 	  void *buffer;
 	  int mpi_tag;
 	  int source;
+	  int req_ready;
+	  starpu_pthread_mutex_t req_mutex;
+	  starpu_pthread_cond_t req_cond;
 );
 
 struct _starpu_mpi_copy_handle_hashlist
@@ -1022,6 +1025,13 @@ static void _starpu_mpi_submit_new_mpi_request(void *arg)
 			 * bring the data back to the original data handle associated to the request.*/
 			if (chandle)
 			{
+				STARPU_PTHREAD_MUTEX_UNLOCK(&mutex);
+				STARPU_PTHREAD_MUTEX_LOCK(&(chandle->req_mutex));
+				while (!(chandle->req_ready))
+					STARPU_PTHREAD_COND_WAIT(&(chandle->req_cond), &(chandle->req_mutex));
+				STARPU_PTHREAD_MUTEX_UNLOCK(&(chandle->req_mutex));
+				STARPU_PTHREAD_MUTEX_LOCK(&mutex);
+
 				_STARPU_MPI_DEBUG(3, "The RECV request %p with tag %d has already been received, copying previously received data into handle's pointer..\n", req, req->mpi_tag);
 				STARPU_ASSERT(req->data_handle != chandle->handle);
 
@@ -1331,8 +1341,10 @@ static void *_starpu_mpi_progress_thread_func(void *arg)
 					data_handle = starpu_data_get_data_handle_from_tag(recv_env->mpi_tag);
 					STARPU_PTHREAD_MUTEX_LOCK(&mutex);
 
-					struct _starpu_mpi_copy_handle* chandle = malloc(sizeof(struct _starpu_mpi_copy_handle));
+					struct _starpu_mpi_copy_handle* chandle = calloc(1, sizeof(struct _starpu_mpi_copy_handle));
 					STARPU_ASSERT(chandle);
+					STARPU_PTHREAD_MUTEX_INIT(&chandle->req_mutex, NULL);
+					STARPU_PTHREAD_COND_INIT(&chandle->req_cond, NULL);
 					chandle->mpi_tag = recv_env->mpi_tag;
 					chandle->env = recv_env;
 					chandle->source = status.MPI_SOURCE;
@@ -1370,6 +1382,11 @@ static void *_starpu_mpi_progress_thread_func(void *arg)
 					while (!(chandle->req->posted))
 					     STARPU_PTHREAD_COND_WAIT(&(chandle->req->posted_cond), &(chandle->req->posted_mutex));
 					STARPU_PTHREAD_MUTEX_UNLOCK(&(chandle->req->posted_mutex));
+
+					STARPU_PTHREAD_MUTEX_LOCK(&chandle->req_mutex);
+					chandle->req_ready = 1;
+					STARPU_PTHREAD_COND_BROADCAST(&chandle->req_cond);
+					STARPU_PTHREAD_MUTEX_UNLOCK(&chandle->req_mutex);
 					STARPU_PTHREAD_MUTEX_LOCK(&mutex);
 				}
 				/* Case : a matching receive has been found for the incoming data, we handle the correct allocation of the pointer associated to