mm: fix assertion mapping->nrpages == 0 in end_writeback()

[pandora-kernel.git] / block / blk.h

diff --git a/block/blk.h b/block/blk.h
index 6126346..d658628 100644 (file)
--- a/block/blk.h
+++ b/block/blk.h
@@ -62,7 +62,28 @@ static inline struct request *__elv_next_request(struct request_queue *q)
                        return rq;
                }
 
-               if (!q->elevator->ops->elevator_dispatch_fn(q, 0))
+               /*
+                * Flush request is running and flush request isn't queueable
+                * in the drive, we can hold the queue till flush request is
+                * finished. Even we don't do this, driver can't dispatch next
+                * requests and will requeue them. And this can improve
+                * throughput too. For example, we have request flush1, write1,
+                * flush 2. flush1 is dispatched, then queue is hold, write1
+                * isn't inserted to queue. After flush1 is finished, flush2
+                * will be dispatched. Since disk cache is already clean,
+                * flush2 will be finished very soon, so looks like flush2 is
+                * folded to flush1.
+                * Since the queue is hold, a flag is set to indicate the queue
+                * should be restarted later. Please see flush_end_io() for
+                * details.
+                */
+               if (q->flush_pending_idx != q->flush_running_idx &&
+                               !queue_flush_queueable(q)) {
+                       q->flush_queue_delayed = 1;
+                       return NULL;
+               }
+               if (test_bit(QUEUE_FLAG_DEAD, &q->queue_flags) ||
+                   !q->elevator->ops->elevator_dispatch_fn(q, 0))
                        return NULL;
        }
 }