/* XXX Could be per-host device, per-module, or even per-connection */
 static DEFINE_SPINLOCK(gb_operations_lock);
 
+static void gb_operation_insert(struct gb_operation *operation)
+{
+       struct gb_connection *connection = operation->connection;
+       struct rb_root *root = &connection->pending;
+       struct rb_node *node = &operation->node;
+       struct rb_node **link = &root->rb_node;
+       struct rb_node *above = NULL;
+       struct gb_operation_msg_hdr *header;
+       __le16 wire_id;
+
+       /*
+        * Assign the operation's id, and store it in the header of
+        * both request and response message headers.
+        */
+       operation->id = gb_connection_operation_id(connection);
+       wire_id = cpu_to_le16(operation->id);
+       header = operation->request->transfer_buffer;
+       header->id = wire_id;
+
+       /* OK, insert the operation into its connection's tree */
+       spin_lock_irq(&gb_operations_lock);
+
+       while (*link) {
+               struct gb_operation *other;
+
+               above = *link;
+               other = rb_entry(above, struct gb_operation, node);
+               header = other->request->transfer_buffer;
+               if (other->id > operation->id)
+                       link = &above->rb_left;
+               else if (other->id < operation->id)
+                       link = &above->rb_right;
+       }
+       rb_link_node(node, above, link);
+       rb_insert_color(node, root);
+
+       spin_unlock_irq(&gb_operations_lock);
+}
+
+static void gb_operation_remove(struct gb_operation *operation)
+{
+       spin_lock_irq(&gb_operations_lock);
+       rb_erase(&operation->node, &operation->connection->pending);
+       spin_unlock_irq(&gb_operations_lock);
+}
+
+static struct gb_operation *
+gb_operation_find(struct gb_connection *connection, u16 id)
+{
+       struct gb_operation *operation;
+       struct rb_node *node;
+       bool found = false;
+
+       spin_lock_irq(&gb_operations_lock);
+       node = connection->pending.rb_node;
+       while (node && !found) {
+               operation = rb_entry(node, struct gb_operation, node);
+               if (operation->id > id)
+                       node = node->rb_left;
+               else if (operation->id < id)
+                       node = node->rb_right;
+               else
+                       found = true;
+       }
+       spin_unlock_irq(&gb_operations_lock);
+
+       return found ? operation : NULL;
+}
+
 /*
  * An operations's response message has arrived.  If no callback was
  * supplied it was submitted for asynchronous completion, so we notify
         * setting the operation id and submitting the gbuf.
         */
        operation->callback = callback;
+       gb_operation_insert(operation);
        ret = greybus_submit_gbuf(operation->request, GFP_KERNEL);
        if (ret)
                return ret;
  */
 static void gb_operation_gbuf_complete(struct gbuf *gbuf)
 {
-       /* TODO */
+       struct gb_operation *operation;
+       struct gb_operation_msg_hdr *header;
+       u16 id;
+
+       /*
+        * This isn't right, but it keeps things balanced until we
+        * can set up operation response handling.
+        */
+       header = gbuf->transfer_buffer;
+       id = le16_to_cpu(header->id);
+       operation = gb_operation_find(gbuf->connection, id);
+       if (operation)
+               gb_operation_remove(operation);
+       else
+               gb_connection_err(gbuf->connection, "operation not found");
 }
 
 /*