[try_put_and_wait] Part 2: Add implementation of try_put_and_wait fea…

…ture for buffering nodes (#1412)

Co-authored-by: Aleksei Fedotov <aleksei.fedotov@intel.com>

include/oneapi/tbb/detail/_flow_graph_cache_impl.h

@@ -98,9 +98,12 @@ class predecessor_cache : public node_cache< sender<T>, M > {
         // Do not work with the passed pointer here as it may not be fully initialized yet
     }
 
-    bool get_item( output_type& v ) {
+private:
+    bool get_item_impl( output_type& v
+                        __TBB_FLOW_GRAPH_METAINFO_ARG(message_metainfo* metainfo_ptr = nullptr) )
+    {
 
-        bool msg = false;
+        bool successful_get = false;
 
         do {
             predecessor_type *src;
@@ -113,19 +116,36 @@ class predecessor_cache : public node_cache< sender<T>, M > {
             }
 
             // Try to get from this sender
-            msg = src->try_get( v );
+#if __TBB_PREVIEW_FLOW_GRAPH_TRY_PUT_AND_WAIT
+            if (metainfo_ptr) {
+                successful_get = src->try_get( v, *metainfo_ptr );
+            } else
+#endif
+            {
+                successful_get = src->try_get( v );
+            }
 
-            if (msg == false) {
+            if (successful_get == false) {
                 // Relinquish ownership of the edge
                 register_successor(*src, *my_owner);
             } else {
                 // Retain ownership of the edge
                 this->add(*src);
             }
-        } while ( msg == false );
-        return msg;
+        } while ( successful_get == false );
+        return successful_get;
+    }
+public:
+    bool get_item( output_type& v ) {
+        return get_item_impl(v);
     }
 
+#if __TBB_PREVIEW_FLOW_GRAPH_TRY_PUT_AND_WAIT
+    bool get_item( output_type& v, message_metainfo& metainfo ) {
+        return get_item_impl(v, &metainfo);
+    }
+#endif
+
     // If we are removing arcs (rf_clear_edges), call clear() rather than reset().
     void reset() {
         for(;;) {
@@ -158,7 +178,7 @@ class reservable_predecessor_cache : public predecessor_cache< T, M > {
     }
 
     bool try_reserve( output_type &v ) {
-        bool msg = false;
+        bool successful_reserve = false;
 
         do {
             predecessor_type* pred = nullptr;
@@ -172,9 +192,9 @@ class reservable_predecessor_cache : public predecessor_cache< T, M > {
             }
 
             // Try to get from this sender
-            msg = pred->try_reserve( v );
+            successful_reserve = pred->try_reserve( v );
 
-            if (msg == false) {
+            if (successful_reserve == false) {
                 typename mutex_type::scoped_lock lock(this->my_mutex);
                 // Relinquish ownership of the edge
                 register_successor( *pred, *this->my_owner );
@@ -183,9 +203,9 @@ class reservable_predecessor_cache : public predecessor_cache< T, M > {
                 // Retain ownership of the edge
                 this->add( *pred);
             }
-        } while ( msg == false );
+        } while ( successful_reserve == false );
 
-        return msg;
+        return successful_reserve;
     }
 
     bool try_release() {
@@ -342,7 +362,7 @@ class broadcast_cache : public successor_cache<T, M> {
     typedef M mutex_type;
     typedef typename successor_cache<T,M>::successors_type successors_type;
 
-    graph_task* try_put_task_impl( const T& t __TBB_FLOW_GRAPH_METAINFO_ARG(const message_metainfo& metainfo)) {
+    graph_task* try_put_task_impl( const T& t __TBB_FLOW_GRAPH_METAINFO_ARG(const message_metainfo& metainfo) ) {
         graph_task * last_task = nullptr;
         typename mutex_type::scoped_lock l(this->my_mutex, /*write=*/true);
         typename successors_type::iterator i = this->my_successors.begin();
@@ -425,11 +445,15 @@ class round_robin_cache : public successor_cache<T, M> {
         return this->my_successors.size();
     }
 
-    graph_task* try_put_task( const T &t ) override {
+private:
+
+    graph_task* try_put_task_impl( const T &t
+                                   __TBB_FLOW_GRAPH_METAINFO_ARG(const message_metainfo& metainfo) )
+    {
         typename mutex_type::scoped_lock l(this->my_mutex, /*write=*/true);
         typename successors_type::iterator i = this->my_successors.begin();
         while ( i != this->my_successors.end() ) {
-            graph_task* new_task = (*i)->try_put_task(t);
+            graph_task* new_task = (*i)->try_put_task(t __TBB_FLOW_GRAPH_METAINFO_ARG(metainfo));
             if ( new_task ) {
                 return new_task;
             } else {
@@ -444,10 +468,14 @@ class round_robin_cache : public successor_cache<T, M> {
         return nullptr;
     }
 
+public:
+    graph_task* try_put_task(const T& t) override {
+        return try_put_task_impl(t __TBB_FLOW_GRAPH_METAINFO_ARG(message_metainfo{}));
+    }
+
 #if __TBB_PREVIEW_FLOW_GRAPH_TRY_PUT_AND_WAIT
-    // TODO: add support for round robin cache
-    graph_task* try_put_task( const T& t, const message_metainfo& ) override {
-        return try_put_task(t);
+    graph_task* try_put_task( const T& t, const message_metainfo& metainfo ) override {
+        return try_put_task_impl(t, metainfo);
     }
 #endif
 };

include/oneapi/tbb/detail/_flow_graph_impl.h

@@ -172,22 +172,44 @@ class trackable_messages_graph_task : public graph_task {
         }
     }
 
+    trackable_messages_graph_task(graph& g, d1::small_object_allocator& allocator,
+                                  node_priority_t node_priority,
+                                  std::forward_list<d1::wait_context_vertex*>&& msg_waiters)
+        : graph_task(g, allocator, node_priority)
+        , my_msg_wait_context_vertices(std::move(msg_waiters))
+    {
+    }
+
     trackable_messages_graph_task(graph& g, d1::small_object_allocator& allocator,
                                   const std::forward_list<d1::wait_context_vertex*>& msg_waiters)
         : trackable_messages_graph_task(g, allocator, no_priority, msg_waiters) {}
 
+    trackable_messages_graph_task(graph& g, d1::small_object_allocator& allocator,
+                                  std::forward_list<d1::wait_context_vertex*>&& msg_waiters)
+        : trackable_messages_graph_task(g, allocator, no_priority, std::move(msg_waiters)) {}
+
     const std::forward_list<d1::wait_context_vertex*> get_msg_wait_context_vertices() const {
         return my_msg_wait_context_vertices;
     }
 
 protected:
     template <typename DerivedType>
     void finalize(const d1::execution_data& ed) {
+        auto wait_context_vertices = std::move(my_msg_wait_context_vertices);
         auto msg_reference_vertices = std::move(my_msg_reference_vertices);
         graph_task::finalize<DerivedType>(ed);
 
-        for (auto& msg_waiter : msg_reference_vertices) {
-            msg_waiter->release(1);
+        // If there is no thread reference vertices associated with the task
+        // then this task was created by transferring the ownership from other metainfo
+        // instance (e.g. while taking from the buffer)
+        if (msg_reference_vertices.empty()) {
+            for (auto& msg_waiter : wait_context_vertices) {
+                msg_waiter->release(1);
+            }
+        } else {
+            for (auto& msg_waiter : msg_reference_vertices) {
+                msg_waiter->release(1);
+            }
         }
     }
 private:

include/oneapi/tbb/detail/_flow_graph_item_buffer_impl.h

@@ -84,6 +84,13 @@ class item_buffer {
         return element(i).item;
     }
 
+#if __TBB_PREVIEW_FLOW_GRAPH_TRY_PUT_AND_WAIT
+    message_metainfo& get_my_metainfo(size_t i) {
+        __TBB_ASSERT(my_item_valid(i), "attempt to get invalid item");
+        return element(i).metainfo;
+    }
+#endif
+
     // may be called with an empty slot or a slot that has already been constructed into.
     void set_my_item(size_t i, const item_type &o
                      __TBB_FLOW_GRAPH_METAINFO_ARG(const message_metainfo& metainfo))
@@ -102,22 +109,46 @@ class item_buffer {
 #endif
     }
 
+#if __TBB_PREVIEW_FLOW_GRAPH_TRY_PUT_AND_WAIT
+    void set_my_item(size_t i, const item_type& o, message_metainfo&& metainfo) {
+        if(element(i).state != no_item) {
+            destroy_item(i);
+        }
+
+        new(&(element(i).item)) item_type(o);
+        new(&element(i).metainfo) message_metainfo(std::move(metainfo));
+        // Skipping the reservation on metainfo.waiters since the ownership
+        // is moving from metainfo to the cache
+        element(i).state = has_item;
+    }
+#endif
+
     // destructively-fetch an object from the buffer
+#if __TBB_PREVIEW_FLOW_GRAPH_TRY_PUT_AND_WAIT
+    void fetch_item(size_t i, item_type& o, message_metainfo& metainfo) {
+        __TBB_ASSERT(my_item_valid(i), "Trying to fetch an empty slot");
+        o = get_my_item(i);  // could have std::move assign semantics
+        metainfo = std::move(get_my_metainfo(i));
+        destroy_item(i);
+    }
+#else
     void fetch_item(size_t i, item_type &o) {
         __TBB_ASSERT(my_item_valid(i), "Trying to fetch an empty slot");
         o = get_my_item(i);  // could have std::move assign semantics
         destroy_item(i);
     }
+#endif
+
 
     // move an existing item from one slot to another.  The moved-to slot must be unoccupied,
     // the moved-from slot must exist and not be reserved.  The after, from will be empty,
     // to will be occupied but not reserved
     void move_item(size_t to, size_t from) {
         __TBB_ASSERT(!my_item_valid(to), "Trying to move to a non-empty slot");
         __TBB_ASSERT(my_item_valid(from), "Trying to move from an empty slot");
-        set_my_item(to, get_my_item(from));   // could have std::move semantics
+        // could have std::move semantics
+        set_my_item(to, get_my_item(from) __TBB_FLOW_GRAPH_METAINFO_ARG(get_my_metainfo(from)));
         destroy_item(from);
-
     }
 
     // put an item in an empty slot.  Return true if successful, else false
@@ -129,12 +160,29 @@ class item_buffer {
         return true;
     }
 
+#if __TBB_PREVIEW_FLOW_GRAPH_TRY_PUT_AND_WAIT
+    template <typename Metainfo>
+    bool place_item(size_t here, const item_type &me, Metainfo&& metainfo) {
+#if !TBB_DEPRECATED_SEQUENCER_DUPLICATES
+        if(my_item_valid(here)) return false;
+#endif
+        set_my_item(here, me, std::forward<Metainfo>(metainfo));
+        return true;
+    }
+#endif
+
     // could be implemented with std::move semantics
     void swap_items(size_t i, size_t j) {
         __TBB_ASSERT(my_item_valid(i) && my_item_valid(j), "attempt to swap invalid item(s)");
         item_type temp = get_my_item(i);
+#if __TBB_PREVIEW_FLOW_GRAPH_TRY_PUT_AND_WAIT
+        message_metainfo temp_metainfo = get_my_metainfo(i);
+        set_my_item(i, get_my_item(j), get_my_metainfo(j));
+        set_my_item(j, temp, temp_metainfo);
+#else
         set_my_item(i, get_my_item(j));
         set_my_item(j, temp);
+#endif
     }
 
     void destroy_item(size_type i) {

include/oneapi/tbb/detail/_flow_graph_join_impl.h

@@ -1190,6 +1190,13 @@
             return op_data.status == SUCCEEDED;
         }
 
+#if __TBB_PREVIEW_FLOW_GRAPH_TRY_PUT_AND_WAIT
+        // TODO: implement try_get with metainfo for join_node
+        bool try_get( output_type &v, message_metainfo& ) override {
+            return try_get(v);
+        }
+#endif
+
     protected:
         void reset_node(reset_flags f) override {
             input_ports_type::reset(f);

include/oneapi/tbb/detail/_flow_graph_node_impl.h

@@ -213,9 +213,12 @@ class function_input_base : public receiver<Input>, no_assign {
         }
         else {
             input_type i;
-            if(my_predecessors.get_item(i)) {
+#if __TBB_PREVIEW_FLOW_GRAPH_TRY_PUT_AND_WAIT
+            message_metainfo metainfo;
+#endif
+            if(my_predecessors.get_item(i __TBB_FLOW_GRAPH_METAINFO_ARG(metainfo))) {
                 ++my_concurrency;
-                new_task = create_body_task(i __TBB_FLOW_GRAPH_METAINFO_ARG(message_metainfo{}));
+                new_task = create_body_task(i __TBB_FLOW_GRAPH_METAINFO_ARG(std::move(metainfo)));
             }
         }
         return new_task;
@@ -351,8 +354,12 @@ class function_input_base : public receiver<Input>, no_assign {
     }
 
     //! allocates a task to apply a body
-    graph_task* create_body_task( const input_type &input
-                                  __TBB_FLOW_GRAPH_METAINFO_ARG(const message_metainfo& metainfo))
+#if __TBB_PREVIEW_FLOW_GRAPH_TRY_PUT_AND_WAIT
+    template <typename Metainfo>
+    graph_task* create_body_task( const input_type &input, Metainfo&& metainfo )
+#else
+    graph_task* create_body_task( const input_type &input )
+#endif
     {
         if (!is_graph_active(my_graph_ref)) {
             return nullptr;
@@ -363,7 +370,7 @@ class function_input_base : public receiver<Input>, no_assign {
 #if __TBB_PREVIEW_FLOW_GRAPH_TRY_PUT_AND_WAIT
         if (!metainfo.empty()) {
             using task_type = apply_body_task_bypass<class_type, input_type, trackable_messages_graph_task>;
-            t = allocator.new_object<task_type>(my_graph_ref, allocator, *this, input, my_priority, metainfo);
+            t = allocator.new_object<task_type>(my_graph_ref, allocator, *this, input, my_priority, std::forward<Metainfo>(metainfo));
         } else
 #endif
         {