Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Peilin Ye <yepeilin.cs@gmail.com>
To:     "David S. Miller" <davem@davemloft.net>,
        Eric Dumazet <edumazet@google.com>,
        Jakub Kicinski <kuba@kernel.org>,
        Paolo Abeni <pabeni@redhat.com>,
        Jonathan Corbet <corbet@lwn.net>,
        Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>,
        David Ahern <dsahern@kernel.org>,
        Jamal Hadi Salim <jhs@mojatatu.com>,
        Cong Wang <xiyou.wangcong@gmail.com>,
        Jiri Pirko <jiri@resnulli.us>
Cc:     Peilin Ye <peilin.ye@bytedance.com>, netdev@vger.kernel.org,
        linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org,
        Cong Wang <cong.wang@bytedance.com>,
        Stephen Hemminger <stephen@networkplumber.org>,
        Dave Taht <dave.taht@gmail.com>,
        Peilin Ye <yepeilin.cs@gmail.com>
Subject: [PATCH RFC v2 net-next 5/5] net/sched: sch_cbq: Use Qdisc backpressure infrastructure
Date:   Mon, 22 Aug 2022 02:12:57 -0700
Message-Id: <614f8f31e3b62dfebb8cb4707c81918a6c7e381d.1661158173.git.peilin.ye@bytedance.com>
In-Reply-To: <cover.1661158173.git.peilin.ye@bytedance.com>
References: <cover.1661158173.git.peilin.ye@bytedance.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: Peilin Ye <peilin.ye@bytedance.com>

Recently we introduced a Qdisc backpressure infrastructure (currently
supports UDP sockets).  Use it in CBQ Qdisc.

Tested with 500 Mbits/sec rate limit using 16 iperf UDP 1 Gbit/sec
clients.  Before:

[  3]  0.0-15.0 sec  55.8 MBytes  31.2 Mbits/sec   1.185 ms 1073326/1113110 (96%)
[  3]  0.0-15.0 sec  55.9 MBytes  31.3 Mbits/sec   1.001 ms 1080330/1120201 (96%)
[  3]  0.0-15.0 sec  55.6 MBytes  31.1 Mbits/sec   1.750 ms 1078292/1117980 (96%)
[  3]  0.0-15.0 sec  55.3 MBytes  30.9 Mbits/sec   0.895 ms 1089200/1128640 (97%)
<...>                                                       ^^^^^^^^^^^^^^^^^^^^^

Total throughput is 493.7 Mbits/sec and average drop rate is 96.13%.

Now enable Qdisc backpressure for UDP sockets, with
udp_backpressure_interval default to 100 milliseconds:

[  3]  0.0-15.0 sec  54.2 MBytes  30.3 Mbits/sec   2.302 ms 54/38692 (0.14%)
[  3]  0.0-15.0 sec  54.1 MBytes  30.2 Mbits/sec   2.227 ms 54/38671 (0.14%)
[  3]  0.0-15.0 sec  53.5 MBytes  29.9 Mbits/sec   2.043 ms 57/38203 (0.15%)
[  3]  0.0-15.0 sec  58.1 MBytes  32.5 Mbits/sec   1.843 ms 1/41480 (0.0024%)
<...>                                                       ^^^^^^^^^^^^^^^^^

Total throughput is 497.1 Mbits/sec (0.69% higher), average drop rate is
0.08% (99.9% lower).

Fairness between flows is slightly affected, with per-flow average
throughput ranging from 29.9 to 32.6 Mbits/sec (compared with 30.3 to
31.3 Mbits/sec).

Signed-off-by: Peilin Ye <peilin.ye@bytedance.com>
---
 net/sched/sch_cbq.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c
index 91a0dc463c48..42e44f570988 100644
--- a/net/sched/sch_cbq.c
+++ b/net/sched/sch_cbq.c
@@ -381,6 +381,7 @@ cbq_enqueue(struct sk_buff *skb, struct Qdisc *sch,
 		return ret;
 	}
 
+	qdisc_backpressure(skb);
 	if (net_xmit_drop_count(ret)) {
 		qdisc_qstats_drop(sch);
 		cbq_mark_toplevel(q, cl);
-- 
2.20.1