From: Jeff Moyer <jmoyer@redhat.com>
Date: Tue, 3 Nov 2009 11:36:45 -0500
Subject: [block] cfq-iosched: fix idling interfering with plugging
Message-id: 1257266206-24003-12-git-send-email-jmoyer@redhat.com
O-Subject: [PATCH 11/12] cfq-iosched: don't let idling interfere with plugging
Bugzilla: 456181 448130 427709
RH-Acked-by: Josef Bacik <josef@redhat.com>
RH-Acked-by: Rik van Riel <riel@redhat.com>
RH-Acked-by: Vivek Goyal <vgoyal@redhat.com>

This commit is necessary to get any semblance of reasonable sequential
read throughput.  It should help workloads other than just the close
cooperators.

commit b029195dda0129b427c6e579a3bb3ae752da3a93
Author: Jens Axboe <jens.axboe@oracle.com>
Date:   Tue Apr 7 11:38:31 2009 +0200

    cfq-iosched: don't let idling interfere with plugging

    When CFQ is waiting for a new request from a process, currently it'll
    immediately restart queuing when it sees such a request. This doesn't
    work very well with streamed IO, since we then end up splitting IO
    that would otherwise have been merged nicely. For a simple dd test,
    this causes 10x as many requests to be issued as we should have.
    Normally this goes unnoticed due to the low overhead of requests
    at the device side, but some hardware is very sensitive to request
    sizes and there it can cause big slow downs.

    Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Jeff Moyer <jmoyer@redhat.com>

diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
index 08275c1..b95f605 100644
--- a/block/cfq-iosched.c
+++ b/block/cfq-iosched.c
@@ -854,6 +854,7 @@ __cfq_set_active_queue(struct cfq_data *cfqd, struct cfq_queue *cfqq)
 		cfqq->slice_start = jiffies;
 		cfqq->slice_end = 0;
 		cfqq->slice_left = 0;
+		cfq_clear_cfqq_must_dispatch(cfqq);
 		cfq_clear_cfqq_must_alloc_slice(cfqq);
 		cfq_clear_cfqq_fifo_expire(cfqq);
 	}
@@ -2102,15 +2103,13 @@ cfq_crq_enqueued(struct cfq_data *cfqd, struct cfq_queue *cfqq,
 
 	if (cfqq == cfqd->active_queue) {
 		/*
-		 * if we are waiting for a request for this queue, let it rip
-		 * immediately and flag that we must not expire this queue
-		 * just now
+		 * Remember that we saw a request from this process, but
+		 * don't start queuing just yet. Otherwise we risk seeing lots
+		 * of tiny requests, because we disrupt the normal plugging
+		 * and merging.
 		 */
-		if (cfq_cfqq_wait_request(cfqq)) {
+		if (cfq_cfqq_wait_request(cfqq))
 			cfq_mark_cfqq_must_dispatch(cfqq);
-			del_timer(&cfqd->idle_slice_timer);
-			cfq_start_queueing(cfqd, cfqq);
-		}
 	} else if (cfq_should_preempt(cfqd, cfqq, crq)) {
 		/*
 		 * not the active queue - expire current slice if it is