--- /dev/null
+From: Eric Dumazet <edumazet@google.com>
+Date: Sun, 1 May 2016 16:47:26 -0700
+Subject: [PATCH] fq_codel: add batch ability to fq_codel_drop()
+
+In presence of inelastic flows and stress, we can call
+fq_codel_drop() for every packet entering fq_codel qdisc.
+
+fq_codel_drop() is quite expensive, as it does a linear scan
+of 4 KB of memory to find a fat flow.
+Once found, it drops the oldest packet of this flow.
+
+Instead of dropping a single packet, try to drop 50% of the backlog
+of this fat flow, with a configurable limit of 64 packets per round.
+
+TCA_FQ_CODEL_DROP_BATCH_SIZE is the new attribute to make this
+limit configurable.
+
+With this strategy the 4 KB search is amortized to a single cache line
+per drop [1], so fq_codel_drop() no longer appears at the top of kernel
+profile in presence of few inelastic flows.
+
+[1] Assuming a 64byte cache line, and 1024 buckets
+
+Signed-off-by: Eric Dumazet <edumazet@google.com>
+Reported-by: Dave Taht <dave.taht@gmail.com>
+Cc: Jonathan Morton <chromatix99@gmail.com>
+Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
+Acked-by: Dave Taht
+Signed-off-by: David S. Miller <davem@davemloft.net>
+---
+
+--- a/include/uapi/linux/pkt_sched.h
++++ b/include/uapi/linux/pkt_sched.h
+@@ -711,6 +711,7 @@ enum {
+ TCA_FQ_CODEL_FLOWS,
+ TCA_FQ_CODEL_QUANTUM,
+ TCA_FQ_CODEL_CE_THRESHOLD,
++ TCA_FQ_CODEL_DROP_BATCH_SIZE,
+ __TCA_FQ_CODEL_MAX
+ };
+
+--- a/net/sched/sch_fq_codel.c
++++ b/net/sched/sch_fq_codel.c
+@@ -57,6 +57,7 @@ struct fq_codel_sched_data {
+ u32 flows_cnt; /* number of flows */
+ u32 perturbation; /* hash perturbation */
+ u32 quantum; /* psched_mtu(qdisc_dev(sch)); */
++ u32 drop_batch_size;
+ struct codel_params cparams;
+ struct codel_stats cstats;
+ u32 drop_overlimit;
+@@ -133,17 +134,20 @@ static inline void flow_queue_add(struct
+ skb->next = NULL;
+ }
+
+-static unsigned int fq_codel_drop(struct Qdisc *sch)
++static unsigned int fq_codel_drop(struct Qdisc *sch, unsigned int max_packets)
+ {
+ struct fq_codel_sched_data *q = qdisc_priv(sch);
+ struct sk_buff *skb;
+ unsigned int maxbacklog = 0, idx = 0, i, len;
+ struct fq_codel_flow *flow;
++ unsigned int threshold;
+
+- /* Queue is full! Find the fat flow and drop packet from it.
++ /* Queue is full! Find the fat flow and drop packet(s) from it.
+ * This might sound expensive, but with 1024 flows, we scan
+ * 4KB of memory, and we dont need to handle a complex tree
+ * in fast path (packet queue/enqueue) with many cache misses.
++ * In stress mode, we'll try to drop 64 packets from the flow,
++ * amortizing this linear lookup to one cache line per drop.
+ */
+ for (i = 0; i < q->flows_cnt; i++) {
+ if (q->backlogs[i] > maxbacklog) {
+@@ -151,15 +155,24 @@ static unsigned int fq_codel_drop(struct
+ idx = i;
+ }
+ }
++
++ /* Our goal is to drop half of this fat flow backlog */
++ threshold = maxbacklog >> 1;
++
+ flow = &q->flows[idx];
+- skb = dequeue_head(flow);
+- len = qdisc_pkt_len(skb);
++ len = 0;
++ i = 0;
++ do {
++ skb = dequeue_head(flow);
++ len += qdisc_pkt_len(skb);
++ kfree_skb(skb);
++ } while (++i < max_packets && len < threshold);
++
++ flow->dropped += i;
+ q->backlogs[idx] -= len;
+- sch->q.qlen--;
+- qdisc_qstats_drop(sch);
+- qdisc_qstats_backlog_dec(sch, skb);
+- kfree_skb(skb);
+- flow->dropped++;
++ sch->qstats.drops += i;
++ sch->qstats.backlog -= len;
++ sch->q.qlen -= i;
+ return idx;
+ }
+
+@@ -168,14 +181,14 @@ static unsigned int fq_codel_qdisc_drop(
+ unsigned int prev_backlog;
+
+ prev_backlog = sch->qstats.backlog;
+- fq_codel_drop(sch);
++ fq_codel_drop(sch, 1U);
+ return prev_backlog - sch->qstats.backlog;
+ }
+
+ static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+ {
+ struct fq_codel_sched_data *q = qdisc_priv(sch);
+- unsigned int idx, prev_backlog;
++ unsigned int idx, prev_backlog, prev_qlen;
+ struct fq_codel_flow *flow;
+ int uninitialized_var(ret);
+
+@@ -204,16 +217,22 @@ static int fq_codel_enqueue(struct sk_bu
+ return NET_XMIT_SUCCESS;
+
+ prev_backlog = sch->qstats.backlog;
+- q->drop_overlimit++;
+- /* Return Congestion Notification only if we dropped a packet
+- * from this flow.
++ prev_qlen = sch->q.qlen;
++
++ /* fq_codel_drop() is quite expensive, as it performs a linear search
++ * in q->backlogs[] to find a fat flow.
++ * So instead of dropping a single packet, drop half of its backlog
++ * with a 64 packets limit to not add a too big cpu spike here.
+ */
+- if (fq_codel_drop(sch) == idx)
+- return NET_XMIT_CN;
++ ret = fq_codel_drop(sch, q->drop_batch_size);
++
++ q->drop_overlimit += prev_qlen - sch->q.qlen;
+
+- /* As we dropped a packet, better let upper stack know this */
+- qdisc_tree_reduce_backlog(sch, 1, prev_backlog - sch->qstats.backlog);
+- return NET_XMIT_SUCCESS;
++ /* As we dropped packet(s), better let upper stack know this */
++ qdisc_tree_reduce_backlog(sch, prev_qlen - sch->q.qlen,
++ prev_backlog - sch->qstats.backlog);
++
++ return ret == idx ? NET_XMIT_CN : NET_XMIT_SUCCESS;
+ }
+
+ /* This is the specific function called from codel_dequeue()
+@@ -323,6 +342,7 @@ static const struct nla_policy fq_codel_
+ [TCA_FQ_CODEL_FLOWS] = { .type = NLA_U32 },
+ [TCA_FQ_CODEL_QUANTUM] = { .type = NLA_U32 },
+ [TCA_FQ_CODEL_CE_THRESHOLD] = { .type = NLA_U32 },
++ [TCA_FQ_CODEL_DROP_BATCH_SIZE] = { .type = NLA_U32 },
+ };
+
+ static int fq_codel_change(struct Qdisc *sch, struct nlattr *opt)
+@@ -374,6 +394,9 @@ static int fq_codel_change(struct Qdisc
+ if (tb[TCA_FQ_CODEL_QUANTUM])
+ q->quantum = max(256U, nla_get_u32(tb[TCA_FQ_CODEL_QUANTUM]));
+
++ if (tb[TCA_FQ_CODEL_DROP_BATCH_SIZE])
++ q->drop_batch_size = min(1U, nla_get_u32(tb[TCA_FQ_CODEL_DROP_BATCH_SIZE]));
++
+ while (sch->q.qlen > sch->limit) {
+ struct sk_buff *skb = fq_codel_dequeue(sch);
+
+@@ -419,6 +442,7 @@ static int fq_codel_init(struct Qdisc *s
+
+ sch->limit = 10*1024;
+ q->flows_cnt = 1024;
++ q->drop_batch_size = 64;
+ q->quantum = psched_mtu(qdisc_dev(sch));
+ q->perturbation = prandom_u32();
+ INIT_LIST_HEAD(&q->new_flows);
+@@ -476,6 +500,8 @@ static int fq_codel_dump(struct Qdisc *s
+ q->cparams.ecn) ||
+ nla_put_u32(skb, TCA_FQ_CODEL_QUANTUM,
+ q->quantum) ||
++ nla_put_u32(skb, TCA_FQ_CODEL_DROP_BATCH_SIZE,
++ q->drop_batch_size) ||
+ nla_put_u32(skb, TCA_FQ_CODEL_FLOWS,
+ q->flows_cnt))
+ goto nla_put_failure;
projects / openwrt / openwrt.git / blobdiff