Commit 276a98a4 authored by Eric Dumazet's avatar Eric Dumazet Committed by Jakub Kicinski
Browse files

net/sched: sch_cake: annotate data-races in cake_dump_stats() (III) 



cake_dump_stats() runs without qdisc spinlock being held.

In this third patch, I add READ_ONCE()/WRITE_ONCE() annotations
for the following fields:

- packets
- tin_dropped
- tin_ecn_mark
- ack_drops
- peak_delay
- avge_delay
- base_delay

Other annotations are added in following patches, to ease code review.

Fixes: 046f6fd5 ("sched: Add Common Applications Kept Enhanced (cake) qdisc")
Signed-off-by: default avatarEric Dumazet <edumazet@google.com>
Acked-by: default avatar"Toke Høiland-Jørgensen" <toke@toke.dk>
Link: https://patch.msgid.link/20260427083606.459355-4-edumazet@google.com


Signed-off-by: default avatarJakub Kicinski <kuba@kernel.org>
parent 91a96427

net/sched/sch_cake.c

+20 −18
Original line number Diff line number Diff line
@@ -1600,7 +1600,7 @@ static unsigned int cake_drop(struct Qdisc *sch, struct sk_buff **to_free) 
	sch->qstats.backlog -= len;



	flow->dropped++;

	b->tin_dropped++;

	WRITE_ONCE(b->tin_dropped, b->tin_dropped + 1);



	if (q->config->rate_flags & CAKE_FLAG_INGRESS)

		cake_advance_shaper(q, b, skb, now, true);
@@ -1820,7 +1820,7 @@ static s32 cake_enqueue(struct sk_buff *skb, struct Qdisc *sch, 
			numsegs++;

			slen += segs->len;

			q->buffer_used += segs->truesize;

			b->packets++;

			WRITE_ONCE(b->packets, b->packets + 1);

		}



		/* stats */
@@ -1844,7 +1844,7 @@ static s32 cake_enqueue(struct sk_buff *skb, struct Qdisc *sch, 
			ack = cake_ack_filter(q, flow);



		if (ack) {

			b->ack_drops++;

			WRITE_ONCE(b->ack_drops, b->ack_drops + 1);

			sch->qstats.drops++;

			ack_pkt_len = qdisc_pkt_len(ack);

			b->bytes += ack_pkt_len;
@@ -1860,7 +1860,7 @@ static s32 cake_enqueue(struct sk_buff *skb, struct Qdisc *sch, 
		}



		/* stats */

		b->packets++;

		WRITE_ONCE(b->packets, b->packets + 1);

		b->bytes	    += len - ack_pkt_len;

		b->backlogs[idx]    += len - ack_pkt_len;

		b->tin_backlog      += len - ack_pkt_len;
@@ -2236,7 +2236,7 @@ static struct sk_buff *cake_dequeue(struct Qdisc *sch) 
			b->tin_deficit -= len;

		}

		flow->dropped++;

		b->tin_dropped++;

		WRITE_ONCE(b->tin_dropped, b->tin_dropped + 1);

		qdisc_tree_reduce_backlog(sch, 1, qdisc_pkt_len(skb));

		qdisc_qstats_drop(sch);

		qdisc_dequeue_drop(sch, skb, reason);
@@ -2244,17 +2244,19 @@ static struct sk_buff *cake_dequeue(struct Qdisc *sch) 
			goto retry;

	}



	b->tin_ecn_mark += !!flow->cvars.ecn_marked;

	WRITE_ONCE(b->tin_ecn_mark, b->tin_ecn_mark + !!flow->cvars.ecn_marked);

	qdisc_bstats_update(sch, skb);

	WRITE_ONCE(q->last_active, now);



	/* collect delay stats */

	delay = ktime_to_ns(ktime_sub(now, cobalt_get_enqueue_time(skb)));

	b->avge_delay = cake_ewma(b->avge_delay, delay, 8);

	b->peak_delay = cake_ewma(b->peak_delay, delay,

				  delay > b->peak_delay ? 2 : 8);

	b->base_delay = cake_ewma(b->base_delay, delay,

				  delay < b->base_delay ? 2 : 8);

	WRITE_ONCE(b->avge_delay, cake_ewma(b->avge_delay, delay, 8));

	WRITE_ONCE(b->peak_delay,

		   cake_ewma(b->peak_delay, delay,

			     delay > b->peak_delay ? 2 : 8));

	WRITE_ONCE(b->base_delay,

		   cake_ewma(b->base_delay, delay,

			     delay < b->base_delay ? 2 : 8));



	len = cake_advance_shaper(q, b, skb, now, false);

	flow->deficit -= len;
@@ -3042,17 +3044,17 @@ static int cake_dump_stats(struct Qdisc *sch, struct gnet_dump *d) 
		PUT_TSTAT_U32(INTERVAL_US,

			      ktime_to_us(ns_to_ktime(b->cparams.interval)));



		PUT_TSTAT_U32(SENT_PACKETS, b->packets);

		PUT_TSTAT_U32(DROPPED_PACKETS, b->tin_dropped);

		PUT_TSTAT_U32(ECN_MARKED_PACKETS, b->tin_ecn_mark);

		PUT_TSTAT_U32(ACKS_DROPPED_PACKETS, b->ack_drops);

		PUT_TSTAT_U32(SENT_PACKETS, READ_ONCE(b->packets));

		PUT_TSTAT_U32(DROPPED_PACKETS, READ_ONCE(b->tin_dropped));

		PUT_TSTAT_U32(ECN_MARKED_PACKETS, READ_ONCE(b->tin_ecn_mark));

		PUT_TSTAT_U32(ACKS_DROPPED_PACKETS, READ_ONCE(b->ack_drops));



		PUT_TSTAT_U32(PEAK_DELAY_US,

			      ktime_to_us(ns_to_ktime(b->peak_delay)));

			      ktime_to_us(ns_to_ktime(READ_ONCE(b->peak_delay))));

		PUT_TSTAT_U32(AVG_DELAY_US,

			      ktime_to_us(ns_to_ktime(b->avge_delay)));

			      ktime_to_us(ns_to_ktime(READ_ONCE(b->avge_delay))));

		PUT_TSTAT_U32(BASE_DELAY_US,

			      ktime_to_us(ns_to_ktime(b->base_delay)));

			      ktime_to_us(ns_to_ktime(READ_ONCE(b->base_delay))));



		PUT_TSTAT_U32(WAY_INDIRECT_HITS, READ_ONCE(b->way_hits));

		PUT_TSTAT_U32(WAY_MISSES, READ_ONCE(b->way_misses));