[PATCH 5/9]: sch_netem: Use requeue list instead of ops->requeue()

[PATCH 5/9]: sch_netem: Use requeue list instead of ops->requeue()

[David Miller]

sch_netem: Use requeue list instead of ops->requeue()

This code just wants to make this packet the "front" one, and that's
just as simply done by queueing to the ->requeue list.

Signed-off-by: David S. Miller <davem@davemloft.net>
---
 net/sched/sch_netem.c |   11 +++--------
 1 files changed, 3 insertions(+), 8 deletions(-)

diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c
index dbd567f..7e15c0b 100644
--- a/net/sched/sch_netem.c
+++ b/net/sched/sch_netem.c
@@ -233,7 +233,8 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 		 */
 		cb->time_to_send = psched_get_time();
 		q->counter = 0;
-		ret = q->qdisc->ops->requeue(skb, q->qdisc);
+		__skb_queue_tail(&q->qdisc->requeue, skb);
+		ret = NET_XMIT_SUCCESS;
 	}
 
 	if (likely(ret == NET_XMIT_SUCCESS)) {
@@ -295,13 +296,7 @@ static struct sk_buff *netem_dequeue(struct Qdisc *sch)
 			return skb;
 		}
 
-		if (unlikely(q->qdisc->ops->requeue(skb, q->qdisc) != NET_XMIT_SUCCESS)) {
-			qdisc_tree_decrease_qlen(q->qdisc, 1);
-			sch->qstats.drops++;
-			printk(KERN_ERR "netem: %s could not requeue\n",
-			       q->qdisc->ops->id);
-		}
-
+		__skb_queue_tail(&q->qdisc->requeue, skb);
 		qdisc_watchdog_schedule(&q->watchdog, cb->time_to_send);
 	}
 
-- 
1.5.6.5.GIT

Re: [PATCH 5/9]: sch_netem: Use requeue list instead of ops->requeue()

[Bastian Bloessl]

David Miller wrote:
> sch_netem: Use requeue list instead of ops->requeue()
> 
> This code just wants to make this packet the "front" one, and that's
> just as simply done by queueing to the ->requeue list.
> 

Hi,
I think that might make some difference.
Suppose we have an upper qdisc that dequeues from netem and
puts the packet in it's requeue-list. But after that another
packet could arrive that should be send earlier (delay
distribution, reordering).
As far as I understand it the requeued packet would now
be send before the other.

> --- a/net/sched/sch_netem.c
> +++ b/net/sched/sch_netem.c
> @@ -233,7 +233,8 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch)
>  		 */
>  		cb->time_to_send = psched_get_time();
>  		q->counter = 0;
> -		ret = q->qdisc->ops->requeue(skb, q->qdisc);
> +		__skb_queue_tail(&q->qdisc->requeue, skb);

I think this should be __skb_queue_head() because in it's
requeue-list netem has packets that it doesn't want to send yet. So
their time_to_send is likely to be in the future. But reordered
packets have current time as time_to_send.

Or did I get it wrong?

Regards
Bastian Bloessl