block: enable rq CPU completion affinity by default

Test results here look good, and on big OLTP runs it has also shown
to significantly increase cycles attributed to the database and
cause a performance boost.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 98b4563..8bf1a10 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -463,7 +463,8 @@
 
 #define QUEUE_FLAG_DEFAULT	((1 << QUEUE_FLAG_IO_STAT) |		\
 				 (1 << QUEUE_FLAG_CLUSTER) |		\
-				 (1 << QUEUE_FLAG_STACKABLE))
+				 (1 << QUEUE_FLAG_STACKABLE)	|	\
+				 (1 << QUEUE_FLAG_SAME_COMP))
 
 static inline int queue_is_locked(struct request_queue *q)
 {