5.4-stable review patch. If anyone has any objections, please let me know.
------------------
From: Eric Dumazet edumazet@google.com
[ Upstream commit b60fa1c5d01a10e358c509b904d4bead6114d593 ]
The introduction of this schedule point was done in commit 2ba2506ca7ca ("[NET]: Add preemption point in qdisc_run") at a time the loop was not bounded.
Then later in commit d5b8aa1d246f ("net_sched: fix dequeuer fairness") we added a limit on the number of packets.
Now is the time to remove the schedule point, since the default limit of 64 packets matches the number of packets a typical NAPI poll can process in a row.
This solves a latency problem for most TCP receivers under moderate load :
1) host receives a packet. NET_RX_SOFTIRQ is raised by NIC hard IRQ handler
2) __do_softirq() does its first loop, handling NET_RX_SOFTIRQ and calling the driver napi->loop() function
3) TCP stores the skb in socket receive queue:
4) TCP calls sk->sk_data_ready() and wakeups a user thread waiting for EPOLLIN (as a result, need_resched() might now be true)
5) TCP cooks an ACK and sends it.
6) qdisc_run() processes one packet from qdisc, and sees need_resched(), this raises NET_TX_SOFTIRQ (even if there are no more packets in the qdisc)
Then we go back to the __do_softirq() in 2), and we see that new softirqs were raised. Since need_resched() is true, we end up waking ksoftirqd in this path :
if (pending) { if (time_before(jiffies, end) && !need_resched() && --max_restart) goto restart;
wakeup_softirqd(); }
So we have many wakeups of ksoftirqd kernel threads, and more calls to qdisc_run() with associated lock overhead.
Note that another way to solve the issue would be to change TCP to first send the ACK packet, then signal the EPOLLIN, but this changes P99 latencies, as sending the ACK packet can add a long delay.
Signed-off-by: Eric Dumazet edumazet@google.com Signed-off-by: David S. Miller davem@davemloft.net Stable-dep-of: 0345552a653c ("net_sched: limit try_bulk_dequeue_skb() batches") Signed-off-by: Sasha Levin sashal@kernel.org --- net/sched/sch_generic.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-)
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 4250f3cf30e72..19fe5078a8d13 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -407,13 +407,8 @@ void __qdisc_run(struct Qdisc *q) int packets;
while (qdisc_restart(q, &packets)) { - /* - * Ordered by possible occurrence: Postpone processing if - * 1. we've exceeded packet quota - * 2. another process needs the CPU; - */ quota -= packets; - if (quota <= 0 || need_resched()) { + if (quota <= 0) { __netif_schedule(q); break; }