lkml.org 
[lkml]   [2009]   [Dec]   [8]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: [PATCH] cfq-iosched: Wait for next request if we are about to expire the queue
Date
Vivek Goyal <vgoyal@redhat.com> writes:

> If there is a sequential reader running in a group, we wait for next request
> to come in that group after slice expiry and once new request is in, we expire
> the queue. Otherwise we delete the group from service tree and group looses
> its fair share.
>
> So far I was marking a queue as wait_busy if it had consumed its slice and
> it was last queue in the group. But this condition did not cover following
> two cases.
>
> - If think time of process is more than slice left, we will not arm the timer
> and queue will be expired.
>
> - If we hit the boundary condition where slice has not expired after request
> completion but almost immediately after (4-5 ns), select_queue() hits and
> by that time slice has expired because jiffies has incremented by one.
>
> Gui was hitting the boundary condition and not getting fairness numbers
> proportional to weight.
>
> This patch puts the checks for above two conditions and improves the fairness
> numbers for sequential workload on rotational media.
>
> Reported-by: Gui Jianfeng <guijianfeng@cn.fujitsu.com>
> Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
> Acked-by: Gui Jianfeng <guijiafneng@cn.fujitsu.com>
> ---
> block/cfq-iosched.c | 30 ++++++++++++++++++++++++++----
> 1 file changed, 26 insertions(+), 4 deletions(-)
>
> Index: linux1/block/cfq-iosched.c
> ===================================================================
> --- linux1.orig/block/cfq-iosched.c 2009-12-07 14:40:03.000000000 -0500
> +++ linux1/block/cfq-iosched.c 2009-12-07 14:40:04.000000000 -0500
> @@ -3251,6 +3251,29 @@ 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;

This really looks like a hack. At the very least, it requires a big
comment above it explaining why it's there. Would it be possible to
detect such a condition in select_queue and handle it there?

Cheers,
Jeff


\
 
 \ /
  Last update: 2009-12-08 18:55    [W:0.073 / U:0.044 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site