selftests/bpf: Add bpf_first scheduler & test

This patch implements the simplest MPTCP scheduler, named bpf_first,
which always picks the first subflow to send data. It's a sample of
MPTCP BPF scheduler implementations.

This patch defines MPTCP_SCHED_TEST macro, a template for all scheduler
tests. Every scheduler is identified by argument name, and use sysctl
to set net.mptcp.scheduler as "bpf_name" to use this sched. Add two
veth net devices to simulate the multiple addresses case. Use 'ip mptcp
endpoint' command to add the new endpoint ADDR2 to PM netlink. Arguments
addr1/add2 means whether the data has been sent on the first/second
subflow or not. Send data and check bytes_sent of 'ss' output after it
using send_data_and_verify().

Using MPTCP_SCHED_TEST macro to add a new test for this bpf_first
scheduler, the arguments "1 0" means data has been only sent on the
first subflow ADDR1. Run this test by RUN_MPTCP_TEST macro.

Signed-off-by: Geliang Tang <[email protected]>
Acked-by: Paolo Abeni <[email protected]>
Reviewed-by: Mat Martineau <[email protected]>
Reviewed-by: Matthieu Baerts (NGI0) <[email protected]>

Author: Geliang Tang

tools/testing/selftests/bpf/prog_tests/mptcp.c

@@ -12,6 +12,7 @@
 #include "mptcpify.skel.h"
 #include "mptcp_subflow.skel.h"
 #include "mptcp_bpf_iters.skel.h"
+#include "mptcp_bpf_first.skel.h"
 
 #define NS_TEST "mptcp_ns"
 #define ADDR_1	"10.0.1.1"
@@ -49,6 +50,7 @@
 #ifndef TCP_CA_NAME_MAX
 #define TCP_CA_NAME_MAX	16
 #endif
+#define MPTCP_SCHED_NAME_MAX	16
 
 static const unsigned int total_bytes = 10 * 1024 * 1024;
 static int duration;
@@ -641,6 +643,46 @@ static void test_default(void)
 	netns_free(netns);
 }
 
+static void test_bpf_sched(struct bpf_object *obj, char *sched,
+			   bool addr1, bool addr2)
+{
+	char bpf_sched[MPTCP_SCHED_NAME_MAX] = "bpf_";
+	struct netns_obj *netns;
+	struct bpf_link *link;
+	struct bpf_map *map;
+
+	if (!ASSERT_LT(strlen(bpf_sched) + strlen(sched),
+		       MPTCP_SCHED_NAME_MAX, "Scheduler name too long"))
+		return;
+
+	map = bpf_object__find_map_by_name(obj, sched);
+	link = bpf_map__attach_struct_ops(map);
+	if (CHECK(!link, sched, "attach_struct_ops: %d\n", errno))
+		return;
+
+	netns = sched_init("subflow", strcat(bpf_sched, sched));
+	if (!netns)
+		goto fail;
+
+	send_data_and_verify(sched, addr1, addr2);
+
+fail:
+	netns_free(netns);
+	bpf_link__destroy(link);
+}
+
+static void test_first(void)
+{
+	struct mptcp_bpf_first *skel;
+
+	skel = mptcp_bpf_first__open_and_load();
+	if (!ASSERT_OK_PTR(skel, "open_and_load: first"))
+		return;
+
+	test_bpf_sched(skel->obj, "first", WITH_DATA, WITHOUT_DATA);
+	mptcp_bpf_first__destroy(skel);
+}
+
 void test_mptcp(void)
 {
 	if (test__start_subtest("base"))
@@ -653,4 +695,6 @@ void test_mptcp(void)
 		test_iters_subflow();
 	if (test__start_subtest("default"))
 		test_default();
+	if (test__start_subtest("first"))
+		test_first();
 }

tools/testing/selftests/bpf/progs/mptcp_bpf.h

@@ -48,4 +48,10 @@ bpf_mptcp_subflow_ctx(const struct sock *sk) __ksym;
 extern struct sock *
 bpf_mptcp_subflow_tcp_sock(const struct mptcp_subflow_context *subflow) __ksym;
 
+extern void mptcp_subflow_set_scheduled(struct mptcp_subflow_context *subflow,
+					bool scheduled) __ksym;
+
+extern struct mptcp_subflow_context *
+bpf_mptcp_subflow_ctx_by_pos(const struct mptcp_sched_data *data, unsigned int pos) __ksym;
+
 #endif

tools/testing/selftests/bpf/progs/mptcp_bpf_first.c

@@ -0,0 +1,33 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2022, SUSE. */
+
+#include "mptcp_bpf.h"
+#include <bpf/bpf_tracing.h>
+
+char _license[] SEC("license") = "GPL";
+
+SEC("struct_ops")
+void BPF_PROG(mptcp_sched_first_init, struct mptcp_sock *msk)
+{
+}
+
+SEC("struct_ops")
+void BPF_PROG(mptcp_sched_first_release, struct mptcp_sock *msk)
+{
+}
+
+SEC("struct_ops")
+int BPF_PROG(bpf_first_get_subflow, struct mptcp_sock *msk,
+	     struct mptcp_sched_data *data)
+{
+	mptcp_subflow_set_scheduled(bpf_mptcp_subflow_ctx_by_pos(data, 0), true);
+	return 0;
+}
+
+SEC(".struct_ops")
+struct mptcp_sched_ops first = {
+	.init		= (void *)mptcp_sched_first_init,
+	.release	= (void *)mptcp_sched_first_release,
+	.get_subflow	= (void *)bpf_first_get_subflow,
+	.name		= "bpf_first",
+};