net: sched: do not acquire qdisc spinlock in qdisc/class stats dump

Large tc dumps (tc -s {qdisc|class} sh dev ethX) done by Google BwE host
agent [1] are problematic at scale :

For each qdisc/class found in the dump, we currently lock the root qdisc
spinlock in order to get stats. Sampling stats every 5 seconds from
thousands of HTB classes is a challenge when the root qdisc spinlock is
under high pressure. Not only the dumps take time, they also slow
down the fast path (queue/dequeue packets) by 10 % to 20 % in some cases.

An audit of existing qdiscs showed that sch_fq_codel is the only qdisc
that might need the qdisc lock in fq_codel_dump_stats() and
fq_codel_dump_class_stats()

In v2 of this patch, I now use the Qdisc running seqcount to provide
consistent reads of packets/bytes counters, regardless of 32/64 bit arches.

I also changed rate estimators to use the same infrastructure
so that they no longer need to lock root qdisc lock.

[1]
http://static.googleusercontent.com/media/research.google.com/en//pubs/archive/43838.pdf

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Cong Wang <xiyou.wangcong@gmail.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: John Fastabend <john.fastabend@gmail.com>
Cc: Kevin Athey <kda@google.com>
Cc: Xiaotian Pei <xiaotian@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 11 insertions, 10 deletions

diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c
index ddf047df5361..d4a8bbfcc953 100644
--- a/net/sched/sch_api.c
+++ b/net/sched/sch_api.c
@@ -982,7 +982,7 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue,
 			rcu_assign_pointer(sch->stab, stab);
 		}
 		if (tca[TCA_RATE]) {
-			spinlock_t *root_lock;
+			seqcount_t *running;
 
 			err = -EOPNOTSUPP;
 			if (sch->flags & TCQ_F_MQROOT)
@@ -991,14 +991,15 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue,
 			if ((sch->parent != TC_H_ROOT) &&
 			    !(sch->flags & TCQ_F_INGRESS) &&
 			    (!p || !(p->flags & TCQ_F_MQROOT)))
-				root_lock = qdisc_root_sleeping_lock(sch);
+				running = qdisc_root_sleeping_running(sch);
 			else
-				root_lock = qdisc_lock(sch);
+				running = &sch->running;
 
 			err = gen_new_estimator(&sch->bstats,
 						sch->cpu_bstats,
 						&sch->rate_est,
-						root_lock,
+						NULL,
+						running,
 						tca[TCA_RATE]);
 			if (err)
 				goto err_out4;
@@ -1061,7 +1062,8 @@ static int qdisc_change(struct Qdisc *sch, struct nlattr **tca)
 		gen_replace_estimator(&sch->bstats,
 				      sch->cpu_bstats,
 				      &sch->rate_est,
-				      qdisc_root_sleeping_lock(sch),
+				      NULL,
+				      qdisc_root_sleeping_running(sch),
 				      tca[TCA_RATE]);
 	}
 out:
@@ -1369,8 +1371,7 @@ static int tc_fill_qdisc(struct sk_buff *skb, struct Qdisc *q, u32 clid,
 		goto nla_put_failure;
 
 	if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS, TCA_XSTATS,
-					 qdisc_root_sleeping_lock(q), &d,
-					 TCA_PAD) < 0)
+					 NULL, &d, TCA_PAD) < 0)
 		goto nla_put_failure;
 
 	if (q->ops->dump_stats && q->ops->dump_stats(q, &d) < 0)
@@ -1381,7 +1382,8 @@ static int tc_fill_qdisc(struct sk_buff *skb, struct Qdisc *q, u32 clid,
 		cpu_qstats = q->cpu_qstats;
 	}
 
-	if (gnet_stats_copy_basic(&d, cpu_bstats, &q->bstats) < 0 ||
+	if (gnet_stats_copy_basic(qdisc_root_sleeping_running(q),
+				  &d, cpu_bstats, &q->bstats) < 0 ||
 	    gnet_stats_copy_rate_est(&d, &q->bstats, &q->rate_est) < 0 ||
 	    gnet_stats_copy_queue(&d, cpu_qstats, &q->qstats, qlen) < 0)
 		goto nla_put_failure;
@@ -1684,8 +1686,7 @@ static int tc_fill_tclass(struct sk_buff *skb, struct Qdisc *q,
 		goto nla_put_failure;
 
 	if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS, TCA_XSTATS,
-					 qdisc_root_sleeping_lock(q), &d,
-					 TCA_PAD) < 0)
+					 NULL, &d, TCA_PAD) < 0)
 		goto nla_put_failure;
 
 	if (cl_ops->dump_stats && cl_ops->dump_stats(q, cl, &d) < 0)