net_sched: fq_codel: cache skb->truesize into skb->cb

Now we defer skb drops, it makes sense to keep a copy
of skb->truesize in struct codel_skb_cb to avoid one
cache line miss per dropped skb in fq_codel_drop(),
to reduce latencies a bit further.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

2 files changed, 5 insertions, 3 deletions

diff --git a/include/net/codel_qdisc.h b/include/net/codel_qdisc.h
index 8144d9cd2908..098630f83a55 100644
--- a/include/net/codel_qdisc.h
+++ b/include/net/codel_qdisc.h
@@ -52,6 +52,7 @@
 /* Qdiscs using codel plugin must use codel_skb_cb in their own cb[] */
 struct codel_skb_cb {
         codel_time_t enqueue_time;
+        unsigned int mem_usage;
 };
 
 static struct codel_skb_cb *get_codel_cb(const struct sk_buff *skb)
diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c
index f715195459c9..a5ea0e9b6be4 100644
--- a/net/sched/sch_fq_codel.c
+++ b/net/sched/sch_fq_codel.c
@@ -172,7 +172,7 @@ static unsigned int fq_codel_drop(struct Qdisc *sch, unsigned int max_packets,
         do {
                 skb = dequeue_head(flow);
                 len += qdisc_pkt_len(skb);
-                mem += skb->truesize;
+                mem += get_codel_cb(skb)->mem_usage;
                 __qdisc_drop(skb, to_free);
         } while (++i < max_packets && len < threshold);
 
@@ -216,7 +216,8 @@ static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch,
                 flow->deficit = q->quantum;
                 flow->dropped = 0;
         }
-        q->memory_usage += skb->truesize;
+        get_codel_cb(skb)->mem_usage = skb->truesize;
+        q->memory_usage += get_codel_cb(skb)->mem_usage;
         memory_limited = q->memory_usage > q->memory_limit;
         if (++sch->q.qlen <= sch->limit && !memory_limited)
                 return NET_XMIT_SUCCESS;
@@ -267,7 +268,7 @@ static struct sk_buff *dequeue_func(struct codel_vars *vars, void *ctx)
         if (flow->head) {
                 skb = dequeue_head(flow);
                 q->backlogs[flow - q->flows] -= qdisc_pkt_len(skb);
-                q->memory_usage -= skb->truesize;
+                q->memory_usage -= get_codel_cb(skb)->mem_usage;
                 sch->q.qlen--;
                 sch->qstats.backlog -= qdisc_pkt_len(skb);
         }