net_sched: avoid too many hrtimer_start() calls (a9efad8b) · Commits · 戴 / test

I found a serious performance bug in packet schedulers using hrtimers. sch_htb and sch_fq are definitely impacted by this problem. We constantly rearm high resolution timers if some packets are throttled in one (or more) class, and other packets are flying through qdisc on another (non throttled) class. hrtimer_start() does not have the mod_timer() trick of doing nothing if expires value does not change : if (timer_pending(timer) && timer->expires == expires) return 1; This issue is particularly visible when multiple cpus can queue/dequeue packets on the same qdisc, as hrtimer code has to lock a remote base. I used following fix : 1) Change htb to use qdisc_watchdog_schedule_ns() instead of open-coding it. 2) Cache watchdog prior expiration. hrtimer might provide this, but I prefer to not rely on some hrtimer internal. Signed-off-by:

Eric Dumazet <edumazet@google.com> Signed-off-by:

David S. Miller <davem@davemloft.net>

include/net/pkt_sched.h

+1 −0

Original line number	Diff line number	Diff line
		@@ -61,6 +61,7 @@ psched_tdiff_bounded(psched_time_t tv1, psched_time_t tv2, psched_time_t bound)
		}

		struct qdisc_watchdog {
		u64 last_expires;
		struct hrtimer timer;
		struct Qdisc *qdisc;
		};

net/sched/sch_api.c

+4 −0

Original line number	Diff line number	Diff line
		@@ -607,6 +607,10 @@ void qdisc_watchdog_schedule_ns(struct qdisc_watchdog *wd, u64 expires, bool thr
		if (throttle)
		qdisc_throttled(wd->qdisc);

		if (wd->last_expires == expires)
		return;

		wd->last_expires = expires;
		hrtimer_start(&wd->timer,
		ns_to_ktime(expires),
		HRTIMER_MODE_ABS_PINNED);

net/sched/sch_htb.c

+3 −10

Original line number	Diff line number	Diff line
		@@ -928,17 +928,10 @@ ok:
		}
		}
		qdisc_qstats_overlimit(sch);
		if (likely(next_event > q->now)) {
		if (!test_bit(__QDISC_STATE_DEACTIVATED,
		&qdisc_root_sleeping(q->watchdog.qdisc)->state)) {
		ktime_t time = ns_to_ktime(next_event);
		qdisc_throttled(q->watchdog.qdisc);
		hrtimer_start(&q->watchdog.timer, time,
		HRTIMER_MODE_ABS_PINNED);
		}
		} else {
		if (likely(next_event > q->now))
		qdisc_watchdog_schedule_ns(&q->watchdog, next_event, true);
		else
		schedule_work(&q->work);
		}
		fin:
		return skb;
		}

Admin message