Merge branch 'sched-refactor-NOLOCK-qdiscs'

Paolo Abeni says:

====================
sched: refactor NOLOCK qdiscs

With the introduction of NOLOCK qdiscs, pfifo_fast performances in the
uncontended scenario degraded measurably, especially after the commit
eb82a994 ("net: sched, fix OOO packets with pfifo_fast").

This series restore the pfifo_fast performances in such scenario back the
previous level, mainly reducing the number of atomic operations required to
perform the qdisc_run() call. Even performances in the contended scenario
increase measurably.

Note: This series is on top of:

sched: manipulate __QDISC_STATE_RUNNING in qdisc_run_* helpers
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

include/linux/skb_array.h

@@ -97,6 +97,11 @@ static inline bool skb_array_empty_any(struct skb_array *a)
 	return ptr_ring_empty_any(&a->ring);
 }
 
+static inline struct sk_buff *__skb_array_consume(struct skb_array *a)
+{
+	return __ptr_ring_consume(&a->ring);
+}
+
 static inline struct sk_buff *skb_array_consume(struct skb_array *a)
 {
 	return ptr_ring_consume(&a->ring);

include/net/sch_generic.h

@@ -30,7 +30,6 @@ struct qdisc_rate_table {
 enum qdisc_state_t {
 	__QDISC_STATE_SCHED,
 	__QDISC_STATE_DEACTIVATED,
-	__QDISC_STATE_RUNNING,
 };
 
 struct qdisc_size_table {
@@ -102,6 +101,7 @@ struct Qdisc {
 	refcount_t		refcnt;
 
 	spinlock_t		busylock ____cacheline_aligned_in_smp;
+	spinlock_t		seqlock;
 };
 
 static inline void qdisc_refcount_inc(struct Qdisc *qdisc)
@@ -111,17 +111,17 @@ static inline void qdisc_refcount_inc(struct Qdisc *qdisc)
 	refcount_inc(&qdisc->refcnt);
 }
 
-static inline bool qdisc_is_running(const struct Qdisc *qdisc)
+static inline bool qdisc_is_running(struct Qdisc *qdisc)
 {
 	if (qdisc->flags & TCQ_F_NOLOCK)
-		return test_bit(__QDISC_STATE_RUNNING, &qdisc->state);
+		return spin_is_locked(&qdisc->seqlock);
 	return (raw_read_seqcount(&qdisc->running) & 1) ? true : false;
 }
 
 static inline bool qdisc_run_begin(struct Qdisc *qdisc)
 {
 	if (qdisc->flags & TCQ_F_NOLOCK) {
-		if (test_and_set_bit(__QDISC_STATE_RUNNING, &qdisc->state))
+		if (!spin_trylock(&qdisc->seqlock))
 			return false;
 	} else if (qdisc_is_running(qdisc)) {
 		return false;
@@ -138,7 +138,7 @@ static inline void qdisc_run_end(struct Qdisc *qdisc)
 {
 	write_seqcount_end(&qdisc->running);
 	if (qdisc->flags & TCQ_F_NOLOCK)
-		clear_bit(__QDISC_STATE_RUNNING, &qdisc->state);
+		spin_unlock(&qdisc->seqlock);
 }
 
 static inline bool qdisc_may_bulk(const struct Qdisc *qdisc)

net/sched/sch_generic.c

@@ -656,7 +656,7 @@ static struct sk_buff *pfifo_fast_dequeue(struct Qdisc *qdisc)
 		if (__skb_array_empty(q))
 			continue;
 
-		skb = skb_array_consume_bh(q);
+		skb = __skb_array_consume(q);
 	}
 	if (likely(skb)) {
 		qdisc_qstats_cpu_backlog_dec(qdisc, skb);
@@ -697,7 +697,7 @@ static void pfifo_fast_reset(struct Qdisc *qdisc)
 		if (!q->ring.queue)
 			continue;
 
-		while ((skb = skb_array_consume_bh(q)) != NULL)
+		while ((skb = __skb_array_consume(q)) != NULL)
 			kfree_skb(skb);
 	}
 
@@ -858,6 +858,11 @@ struct Qdisc *qdisc_alloc(struct netdev_queue *dev_queue,
 	lockdep_set_class(&sch->busylock,
 			  dev->qdisc_tx_busylock ?: &qdisc_tx_busylock);
 
+	/* seqlock has the same scope of busylock, for NOLOCK qdisc */
+	spin_lock_init(&sch->seqlock);
+	lockdep_set_class(&sch->busylock,
+			  dev->qdisc_tx_busylock ?: &qdisc_tx_busylock);
+
 	seqcount_init(&sch->running);
 	lockdep_set_class(&sch->running,
 			  dev->qdisc_running_key ?: &qdisc_running_key);
@@ -1097,6 +1102,10 @@ static void dev_deactivate_queue(struct net_device *dev,
 
 	qdisc = rtnl_dereference(dev_queue->qdisc);
 	if (qdisc) {
+		bool nolock = qdisc->flags & TCQ_F_NOLOCK;
+
+		if (nolock)
+			spin_lock_bh(&qdisc->seqlock);
 		spin_lock_bh(qdisc_lock(qdisc));
 
 		if (!(qdisc->flags & TCQ_F_BUILTIN))
@@ -1106,6 +1115,8 @@ static void dev_deactivate_queue(struct net_device *dev,
 		qdisc_reset(qdisc);
 
 		spin_unlock_bh(qdisc_lock(qdisc));
+		if (nolock)
+			spin_unlock_bh(&qdisc->seqlock);
 	}
 }