diff options
author | Eric Dumazet <edumazet@google.com> | 2012-05-16 08:39:09 +0400 |
---|---|---|
committer | David S. Miller <davem@davemloft.net> | 2012-05-16 23:30:26 +0400 |
commit | 865ec5523dadbedefbc5710a68969f686a28d928 (patch) | |
tree | b9f056cdea6922d5fc3ed035764660bb6fb152b3 /net/sched/sch_fq_codel.c | |
parent | c27b46e7f1cbf3be95a4cf5840c76a7b7d54b26f (diff) | |
download | linux-865ec5523dadbedefbc5710a68969f686a28d928.tar.xz |
fq_codel: should use qdisc backlog as threshold
codel_should_drop() logic allows a packet being not dropped if queue
size is under max packet size.
In fq_codel, we have two possible backlogs : The qdisc global one, and
the flow local one.
The meaningful one for codel_should_drop() should be the global backlog,
not the per flow one, so that thin flows can have a non zero drop/mark
probability.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Dave Taht <dave.taht@bufferbloat.net>
Cc: Kathleen Nichols <nichols@pollere.com>
Cc: Van Jacobson <van@pollere.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/sched/sch_fq_codel.c')
-rw-r--r-- | net/sched/sch_fq_codel.c | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c index 337ff204f272..9fc1c62ec80e 100644 --- a/net/sched/sch_fq_codel.c +++ b/net/sched/sch_fq_codel.c @@ -217,13 +217,14 @@ static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch) */ static struct sk_buff *dequeue(struct codel_vars *vars, struct Qdisc *sch) { + struct fq_codel_sched_data *q = qdisc_priv(sch); struct fq_codel_flow *flow; struct sk_buff *skb = NULL; flow = container_of(vars, struct fq_codel_flow, cvars); if (flow->head) { skb = dequeue_head(flow); - sch->qstats.backlog -= qdisc_pkt_len(skb); + q->backlogs[flow - q->flows] -= qdisc_pkt_len(skb); sch->q.qlen--; } return skb; @@ -256,7 +257,7 @@ begin: prev_ecn_mark = q->cstats.ecn_mark; skb = codel_dequeue(sch, &q->cparams, &flow->cvars, &q->cstats, - dequeue, &q->backlogs[flow - q->flows]); + dequeue); flow->dropped += q->cstats.drop_count - prev_drop_count; flow->dropped += q->cstats.ecn_mark - prev_ecn_mark; |