From: Vivek Goyal <vgoyal@redhat.com>
To: Gui Jianfeng <guijianfeng@cn.fujitsu.com>
Cc: linux-kernel@vger.kernel.org, jens.axboe@oracle.com,
nauman@google.com, dpshah@google.com, lizf@cn.fujitsu.com,
ryov@valinux.co.jp, fernando@oss.ntt.co.jp,
s-uchida@ap.jp.nec.com, taka@valinux.co.jp, jmoyer@redhat.com,
righi.andrea@gmail.com, m-ikeda@ds.jp.nec.com,
czoccolo@gmail.com, Alan.Brunelle@hp.com
Subject: Re: Block IO Controller V4
Date: Thu, 3 Dec 2009 18:51:53 -0500 [thread overview]
Message-ID: <20091203235153.GG2735@redhat.com> (raw)
In-Reply-To: <20091203181003.GD2735@redhat.com>
On Thu, Dec 03, 2009 at 01:10:03PM -0500, Vivek Goyal wrote:
[..]
> Hi Gui,
>
> Can you please try following patch and see if it helps you. If not, then
> we need to figure out why we choose to not idle and delete the group from
> service tree.
>
Hi Gui,
Please try this version of the patch instead of previous one. During more
testing I saw some additional deletions where we should have waited and
the reason being that we were hitting boundary condition. At the request
completion time slice has not expired but after 4-5 ns, select_queue hits
and jiffy has incremented by then and slice expires.
ttime_mean, is not covering this condition because this workload is so
sequential that ttime_mean=0.
So I am checking for new condition where if we are into last ms of slice,
mark the queue wait_busy.
Thanks
Vivek
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
---
block/cfq-iosched.c | 30 ++++++++++++++++++++++++++----
1 file changed, 26 insertions(+), 4 deletions(-)
Index: linux11/block/cfq-iosched.c
===================================================================
--- linux11.orig/block/cfq-iosched.c 2009-12-03 15:10:53.000000000 -0500
+++ linux11/block/cfq-iosched.c 2009-12-03 18:37:35.000000000 -0500
@@ -3248,6 +3248,29 @@ static void cfq_update_hw_tag(struct cfq
cfqd->hw_tag = 0;
}
+static inline bool
+cfq_should_wait_busy(struct cfq_data *cfqd, struct cfq_queue *cfqq)
+{
+ struct cfq_io_context *cic = cfqd->active_cic;
+
+ /* If there are other queues in the group, don't wait */
+ if (cfqq->cfqg->nr_cfqq > 1)
+ return false;
+
+ if (cfq_slice_used(cfqq))
+ return true;
+
+ if (cfqq->slice_end - jiffies == 1)
+ return true;
+
+ /* if slice left is less than think time, wait busy */
+ if (cic && sample_valid(cic->ttime_samples)
+ && (cfqq->slice_end - jiffies < cic->ttime_mean))
+ return true;
+
+ return false;
+}
+
static void cfq_completed_request(struct request_queue *q, struct request *rq)
{
struct cfq_queue *cfqq = RQ_CFQQ(rq);
@@ -3286,11 +3309,10 @@ static void cfq_completed_request(struct
}
/*
- * If this queue consumed its slice and this is last queue
- * in the group, wait for next request before we expire
- * the queue
+ * Should we wait for next request to come in before we expire
+ * the queue.
*/
- if (cfq_slice_used(cfqq) && cfqq->cfqg->nr_cfqq == 1) {
+ if (cfq_should_wait_busy(cfqd, cfqq)) {
cfqq->slice_end = jiffies + cfqd->cfq_slice_idle;
cfq_mark_cfqq_wait_busy(cfqq);
}
next prev parent reply other threads:[~2009-12-03 23:53 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-11-30 2:59 Block IO Controller V4 Vivek Goyal
2009-11-30 2:59 ` [PATCH 01/21] blkio: Set must_dispatch only if we decided to not dispatch the request Vivek Goyal
2009-12-02 14:06 ` Jeff Moyer
2009-11-30 2:59 ` [PATCH 02/21] blkio: Introduce the notion of cfq groups Vivek Goyal
2009-11-30 2:59 ` [PATCH 03/21] blkio: Implement macro to traverse each idle tree in group Vivek Goyal
2009-11-30 20:13 ` Divyesh Shah
2009-11-30 22:24 ` Vivek Goyal
2009-11-30 2:59 ` [PATCH 04/21] blkio: Keep queue on service tree until we expire it Vivek Goyal
2009-11-30 2:59 ` [PATCH 05/21] blkio: Introduce the root service tree for cfq groups Vivek Goyal
2009-11-30 23:55 ` Divyesh Shah
2009-12-02 15:42 ` Vivek Goyal
2009-12-02 15:49 ` Vivek Goyal
2009-11-30 2:59 ` [PATCH 06/21] blkio: Introduce blkio controller cgroup interface Vivek Goyal
2009-12-01 0:04 ` Divyesh Shah
2009-12-02 15:27 ` Vivek Goyal
2009-11-30 2:59 ` [PATCH 07/21] blkio: Introduce per cfq group weights and vdisktime calculations Vivek Goyal
2009-12-02 15:50 ` Vivek Goyal
2009-11-30 2:59 ` [PATCH 08/21] blkio: Implement per cfq group latency target and busy queue avg Vivek Goyal
2009-11-30 2:59 ` [PATCH 09/21] blkio: Group time used accounting and workload context save restore Vivek Goyal
2009-11-30 2:59 ` [PATCH 10/21] blkio: Dynamic cfq group creation based on cgroup tasks belongs to Vivek Goyal
2009-11-30 2:59 ` [PATCH 11/21] blkio: Take care of cgroup deletion and cfq group reference counting Vivek Goyal
2009-11-30 2:59 ` [PATCH 12/21] blkio: Some debugging aids for CFQ Vivek Goyal
2009-11-30 2:59 ` [PATCH 13/21] blkio: Export disk time and sectors used by a group to user space Vivek Goyal
2009-11-30 2:59 ` [PATCH 14/21] blkio: Provide some isolation between groups Vivek Goyal
2009-11-30 2:59 ` [PATCH 15/21] blkio: Drop the reference to queue once the task changes cgroup Vivek Goyal
2009-11-30 2:59 ` [PATCH 16/21] blkio: Propagate cgroup weight updation to cfq groups Vivek Goyal
2009-11-30 2:59 ` [PATCH 17/21] blkio: Wait for cfq queue to get backlogged if group is empty Vivek Goyal
2009-11-30 2:59 ` [PATCH 18/21] blkio: Determine async workload length based on total number of queues Vivek Goyal
2009-11-30 2:59 ` [PATCH 19/21] blkio: Implement group_isolation tunable Vivek Goyal
2009-11-30 2:59 ` [PATCH 20/21] blkio: Wait on sync-noidle queue even if rq_noidle = 1 Vivek Goyal
2009-11-30 2:59 ` [PATCH 21/21] blkio: Documentation Vivek Goyal
2009-11-30 15:34 ` Block IO Controller V4 Corrado Zoccolo
2009-11-30 16:00 ` Vivek Goyal
2009-11-30 21:34 ` Corrado Zoccolo
2009-11-30 21:58 ` Vivek Goyal
2009-11-30 22:00 ` Alan D. Brunelle
2009-11-30 22:56 ` Vivek Goyal
2009-11-30 23:50 ` Alan D. Brunelle
2009-12-02 19:12 ` Vivek Goyal
2009-12-08 15:17 ` Alan D. Brunelle
2009-12-08 16:32 ` Vivek Goyal
2009-12-08 18:05 ` Alan D. Brunelle
2009-12-10 3:44 ` Vivek Goyal
2009-12-01 22:27 ` Vivek Goyal
2009-12-02 1:51 ` Gui Jianfeng
2009-12-02 14:25 ` Vivek Goyal
2009-12-03 8:41 ` Gui Jianfeng
2009-12-03 14:36 ` Vivek Goyal
2009-12-03 18:10 ` Vivek Goyal
2009-12-03 23:51 ` Vivek Goyal [this message]
2009-12-07 8:45 ` Gui Jianfeng
2009-12-07 15:25 ` Vivek Goyal
2009-12-07 1:35 ` Gui Jianfeng
2009-12-07 8:41 ` Gui Jianfeng
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20091203235153.GG2735@redhat.com \
--to=vgoyal@redhat.com \
--cc=Alan.Brunelle@hp.com \
--cc=czoccolo@gmail.com \
--cc=dpshah@google.com \
--cc=fernando@oss.ntt.co.jp \
--cc=guijianfeng@cn.fujitsu.com \
--cc=jens.axboe@oracle.com \
--cc=jmoyer@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=lizf@cn.fujitsu.com \
--cc=m-ikeda@ds.jp.nec.com \
--cc=nauman@google.com \
--cc=righi.andrea@gmail.com \
--cc=ryov@valinux.co.jp \
--cc=s-uchida@ap.jp.nec.com \
--cc=taka@valinux.co.jp \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox