Message-Id: <1677576294-33411-5-git-send-email-alibuda@linux.alibaba.com>
Date:   Tue, 28 Feb 2023 17:24:54 +0800
From:   "D. Wythe" <alibuda@...ux.alibaba.com>
To:     kgraul@...ux.ibm.com, wenjia@...ux.ibm.com, jaka@...ux.ibm.com,
        ast@...nel.org, daniel@...earbox.net, andrii@...nel.org
Cc:     kuba@...nel.org, davem@...emloft.net, netdev@...r.kernel.org,
        linux-s390@...r.kernel.org, linux-rdma@...r.kernel.org,
        bpf@...r.kernel.org
Subject: [PATCH bpf-next v3 4/4] bpf/selftests: add selftest for SMC bpf capability

From: "D. Wythe" <alibuda@...ux.alibaba.com>

This PATCH adds a tiny selftest for SMC bpf capability,
making decisions on whether to use SMC by collecting
certain information from kernel smc sock.

Follow the steps below to run this test.

make -C tools/testing/selftests/bpf
cd tools/testing/selftests/bpf
sudo ./test_progs -t bpf_smc

Results shows:
18      bpf_smc:OK
Summary: 1/0 PASSED, 0 SKIPPED, 0 FAILED

Signed-off-by: D. Wythe <alibuda@...ux.alibaba.com>
---
 tools/testing/selftests/bpf/prog_tests/bpf_smc.c |  37 +++
 tools/testing/selftests/bpf/progs/bpf_smc.c      | 320 +++++++++++++++++++++++
 2 files changed, 357 insertions(+)
 create mode 100644 tools/testing/selftests/bpf/prog_tests/bpf_smc.c
 create mode 100644 tools/testing/selftests/bpf/progs/bpf_smc.c

diff --git a/tools/testing/selftests/bpf/prog_tests/bpf_smc.c b/tools/testing/selftests/bpf/prog_tests/bpf_smc.c
new file mode 100644
index 0000000..b57326c
--- /dev/null
+++ b/tools/testing/selftests/bpf/prog_tests/bpf_smc.c
@@ -0,0 +1,37 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/err.h>
+#include <netinet/tcp.h>
+#include <test_progs.h>
+#include "bpf_smc.skel.h"
+
+void test_bpf_smc(void)
+{
+	struct bpf_smc *smc_skel;
+	struct bpf_link *link;
+	int err;
+
+	smc_skel = bpf_smc__open();
+	if (!ASSERT_OK_PTR(smc_skel, "skel_open"))
+		return;
+
+	err = bpf_map__set_type(smc_skel->maps.negotiator_map, BPF_MAP_TYPE_HASH);
+	if (!ASSERT_OK(err, "bpf_map__set_type"))
+		goto error;
+
+	err = bpf_map__set_max_entries(smc_skel->maps.negotiator_map, 1);
+	if (!ASSERT_OK(err, "bpf_map__set_type"))
+		goto error;
+
+	err =  bpf_smc__load(smc_skel);
+	if (!ASSERT_OK(err, "skel_load"))
+		goto error;
+
+	link = bpf_map__attach_struct_ops(smc_skel->maps.ops);
+	if (!ASSERT_OK_PTR(link, "bpf_map__attach_struct_ops"))
+		goto error;
+
+	bpf_link__destroy(link);
+error:
+	bpf_smc__destroy(smc_skel);
+}
diff --git a/tools/testing/selftests/bpf/progs/bpf_smc.c b/tools/testing/selftests/bpf/progs/bpf_smc.c
new file mode 100644
index 0000000..4f4a541
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/bpf_smc.c
@@ -0,0 +1,320 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+#include <linux/bpf.h>
+#include <linux/stddef.h>
+#include <linux/smc.h>
+#include <stdbool.h>
+#include <linux/types.h>
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_core_read.h>
+#include <bpf/bpf_tracing.h>
+
+#define AF_SMC			(43)
+#define SMC_LISTEN		(10)
+#define SMC_SOCK_CLOSED_TIMING	(0)
+extern unsigned long CONFIG_HZ __kconfig;
+#define HZ CONFIG_HZ
+
+char _license[] SEC("license") = "GPL";
+#define max(a, b) ((a) > (b) ? (a) : (b))
+
+struct sock_common {
+	unsigned char	skc_state;
+	unsigned short	skc_family;
+	__u16	skc_num;
+} __attribute__((preserve_access_index));
+
+struct sock {
+	struct sock_common	__sk_common;
+	int	sk_sndbuf;
+} __attribute__((preserve_access_index));
+
+struct inet_sock {
+	struct sock	sk;
+} __attribute__((preserve_access_index));
+
+struct inet_connection_sock {
+	struct inet_sock	icsk_inet;
+} __attribute__((preserve_access_index));
+
+struct tcp_sock {
+	struct inet_connection_sock	inet_conn;
+	__u32	rcv_nxt;
+	__u32	snd_nxt;
+	__u32	snd_una;
+	__u32	delivered;
+	__u8	syn_data:1,	/* SYN includes data */
+		syn_fastopen:1,	/* SYN includes Fast Open option */
+		syn_fastopen_exp:1,/* SYN includes Fast Open exp. option */
+		syn_fastopen_ch:1, /* Active TFO re-enabling probe */
+		syn_data_acked:1,/* data in SYN is acked by SYN-ACK */
+		save_syn:1,	/* Save headers of SYN packet */
+		is_cwnd_limited:1,/* forward progress limited by snd_cwnd? */
+		syn_smc:1;	/* SYN includes SMC */
+} __attribute__((preserve_access_index));
+
+struct socket {
+	struct sock *sk;
+} __attribute__((preserve_access_index));
+
+union smc_host_cursor {
+	struct {
+		__u16	reserved;
+		__u16	wrap;
+		__u32	count;
+	};
+} __attribute__((preserve_access_index));
+
+struct smc_connection {
+	union smc_host_cursor	tx_curs_sent;
+	union smc_host_cursor	rx_curs_confirmed;
+} __attribute__((preserve_access_index));
+
+struct smc_sock {
+	struct sock	sk;
+	struct socket	*clcsock;	/* internal tcp socket */
+	struct smc_connection	conn;
+	int use_fallback;
+} __attribute__((preserve_access_index));
+
+static __always_inline struct tcp_sock *tcp_sk(const struct sock *sk)
+{
+	return (struct tcp_sock *)sk;
+}
+
+static __always_inline struct smc_sock *smc_sk(struct sock *sk)
+{
+	return (struct smc_sock *)sk;
+}
+
+struct smc_prediction {
+	/* protection for smc_prediction */
+	struct bpf_spin_lock lock;
+	/* start of time slice */
+	__u64	start_tstamp;
+	/* delta of pacing */
+	__u64	pacing_delta;
+	/* N of closed connections determined as long connections
+	 * in current time slice
+	 */
+	__u32	closed_long_cc;
+	/* N of closed connections in this time slice */
+	__u32	closed_total_cc;
+	/* N of incoming connections determined as long connections
+	 * in current time slice
+	 */
+	__u32	incoming_long_cc;
+	/* last splice rate of long cc */
+	__u32	last_rate_of_lcc;
+};
+
+#define SMC_PREDICTION_MIN_PACING_DELTA                (1llu)
+#define SMC_PREDICTION_MAX_PACING_DELTA                (HZ << 3)
+#define SMC_PREDICTION_MAX_LONGCC_PER_SPLICE           (8)
+#define SMC_PREDICTION_MAX_PORT                        (64)
+#define SMC_PREDICTION_MAX_SPLICE_GAP                  (1)
+#define SMC_PREDICTION_LONGCC_RATE_THRESHOLD           (13189)
+#define SMC_PREDICTION_LONGCC_PACKETS_THRESHOLD        (100)
+#define SMC_PREDICTION_LONGCC_BYTES_THRESHOLD	\
+		(SMC_PREDICTION_LONGCC_PACKETS_THRESHOLD * 1024)
+
+struct {
+	__uint(type, BPF_MAP_TYPE_HASH);
+	__uint(max_entries, SMC_PREDICTION_MAX_PORT);
+	__type(key, __u16);
+	__type(value, struct smc_prediction);
+} negotiator_map SEC(".maps");
+
+
+static inline __u32 smc_prediction_calt_rate(struct smc_prediction *smc_predictor)
+{
+	if (!smc_predictor->closed_total_cc)
+		return smc_predictor->last_rate_of_lcc;
+
+	return (smc_predictor->closed_long_cc << 14) / smc_predictor->closed_total_cc;
+}
+
+static inline struct smc_prediction *smc_prediction_get(__u16 key, __u64 tstamp)
+{
+	struct smc_prediction zero = {}, *smc_predictor;
+	__u32 gap;
+	int err;
+
+	smc_predictor = bpf_map_lookup_elem(&negotiator_map, &key);
+	if (!smc_predictor) {
+		zero.start_tstamp = bpf_jiffies64();
+		zero.pacing_delta = SMC_PREDICTION_MIN_PACING_DELTA;
+		err = bpf_map_update_elem(&negotiator_map, &key, &zero, 0);
+		if (err)
+			return NULL;
+		smc_predictor =  bpf_map_lookup_elem(&negotiator_map, &key);
+		if (!smc_predictor)
+			return NULL;
+	}
+
+	if (tstamp) {
+		bpf_spin_lock(&smc_predictor->lock);
+		gap = (tstamp - smc_predictor->start_tstamp) / smc_predictor->pacing_delta;
+		/* new splice */
+		if (gap > 0) {
+			smc_predictor->start_tstamp = tstamp;
+			smc_predictor->last_rate_of_lcc =
+				(smc_prediction_calt_rate(smc_predictor) * 7) >> (2 + gap);
+			smc_predictor->closed_long_cc = 0;
+			smc_predictor->closed_total_cc = 0;
+			smc_predictor->incoming_long_cc = 0;
+		}
+		bpf_spin_unlock(&smc_predictor->lock);
+	}
+	return smc_predictor;
+}
+
+/* BPF struct ops for smc protocol negotiator */
+struct smc_sock_negotiator_ops {
+	/* ret for negotiate */
+	int (*negotiate)(struct smc_sock *smc);
+
+	/* info gathering timing */
+	void (*collect_info)(struct smc_sock *smc, int timing);
+};
+
+int SEC("struct_ops/bpf_smc_negotiate")
+BPF_PROG(bpf_smc_negotiate, struct smc_sock *smc)
+{
+	struct smc_prediction *smc_predictor;
+	int err, ret = SK_DROP;
+	struct tcp_sock *tp;
+	struct sock *clcsk;
+	__u32 rate = 0;
+	__u16 key;
+
+	/* client side */
+	if (smc == NULL || smc->sk.__sk_common.skc_state != SMC_LISTEN) {
+		/* use Global smc_predictor */
+		key = 0;
+	} else {	/* server side */
+		clcsk = BPF_CORE_READ(smc, clcsock, sk);
+		if (!clcsk)
+			goto error;
+		tp = tcp_sk(clcsk);
+		err = bpf_core_read(&key, sizeof(__u16),
+				    &tp->inet_conn.icsk_inet.sk.__sk_common.skc_num);
+		if (err)
+			goto error;
+	}
+
+	smc_predictor = smc_prediction_get(key, bpf_jiffies64());
+	if (!smc_predictor)
+		return SK_PASS;
+
+	bpf_spin_lock(&smc_predictor->lock);
+
+	if (smc_predictor->incoming_long_cc == 0)
+		goto out_locked_pass;
+
+	if (smc_predictor->incoming_long_cc > SMC_PREDICTION_MAX_LONGCC_PER_SPLICE)
+		goto out_locked_drop;
+
+	rate = smc_prediction_calt_rate(smc_predictor);
+	if (rate < SMC_PREDICTION_LONGCC_RATE_THRESHOLD)
+		goto out_locked_drop;
+
+out_locked_pass:
+	smc_predictor->incoming_long_cc++;
+	bpf_spin_unlock(&smc_predictor->lock);
+	return SK_PASS;
+out_locked_drop:
+	bpf_spin_unlock(&smc_predictor->lock);
+error:
+	return SK_DROP;
+}
+
+void SEC("struct_ops/bpf_smc_collect_info")
+BPF_PROG(bpf_smc_collect_info, struct sock *sk, int timing)
+{
+	struct smc_prediction *smc_predictor;
+	int use_fallback, sndbuf, err;
+	struct smc_sock *smc;
+	struct tcp_sock *tp;
+	struct sock *clcsk;
+	bool match = false;
+	__u16 wrap, count;
+	__u32 delivered;
+	__u16 key;
+
+	/* no info can collect */
+	if (sk == NULL)
+		return;
+
+	/* only fouces on closed */
+	if (timing != SMC_SOCK_CLOSED_TIMING)
+		return;
+
+	/* first check the sk type */
+	if (sk->__sk_common.skc_family == AF_SMC) {
+		smc = smc_sk(sk);
+		clcsk = BPF_CORE_READ(smc, clcsock, sk);
+		if (!clcsk)
+			goto error;
+		tp = tcp_sk(clcsk);
+		/* check if it's fallback */
+		err = bpf_core_read(&use_fallback, sizeof(use_fallback), &smc->use_fallback);
+		if (err)
+			goto error;
+		if (use_fallback)
+			goto fallback;
+		err = bpf_core_read(&wrap, sizeof(__u16), &smc->conn.tx_curs_sent.wrap);
+		if (err)
+			goto error;
+		err = bpf_core_read(&count, sizeof(__u16), &smc->conn.tx_curs_sent.count);
+		if (err)
+			goto error;
+		err = bpf_core_read(&sndbuf, sizeof(int), &clcsk->sk_sndbuf);
+		if (err)
+			goto error;
+		 match = (count + wrap * sndbuf) > SMC_PREDICTION_LONGCC_BYTES_THRESHOLD;
+	} else {
+		smc = NULL;
+		tp = tcp_sk(sk);
+		use_fallback = 1;
+fallback:
+		err = bpf_core_read(&delivered, sizeof(delivered), &tp->delivered);
+		if (err)
+			goto error;
+		match = (delivered > SMC_PREDICTION_LONGCC_PACKETS_THRESHOLD);
+	}
+
+	/* whatever, tp is never NULL */
+	err = bpf_core_read(&key, sizeof(__u16), &tp->inet_conn.icsk_inet.sk.__sk_common.skc_num);
+	if (err)
+		goto error;
+
+	smc_predictor = smc_prediction_get(key, 0);
+	if (!smc_predictor)
+		goto error;
+
+	bpf_spin_lock(&smc_predictor->lock);
+	smc_predictor->closed_total_cc++;
+	if (match) {
+		/* increase stats */
+		smc_predictor->closed_long_cc++;
+		/* try more aggressive */
+		if (smc_predictor->pacing_delta > SMC_PREDICTION_MIN_PACING_DELTA) {
+			if (use_fallback) {
+				smc_predictor->pacing_delta = max(SMC_PREDICTION_MIN_PACING_DELTA,
+						(smc_predictor->pacing_delta * 3) >> 2);
+			}
+		}
+	} else if (!use_fallback) {
+		smc_predictor->pacing_delta <<= 1;
+	}
+	bpf_spin_unlock(&smc_predictor->lock);
+error:
+	return;
+}
+
+SEC(".struct_ops")
+struct smc_sock_negotiator_ops ops = {
+	.negotiate	= (void *)bpf_smc_negotiate,
+	.collect_info	= (void *)bpf_smc_collect_info,
+};
-- 
1.8.3.1

Powered by blists - more mailing lists