selftests: net: test that listening sockets match on address properly

This patch adds a selftest that verifies that a socket listening
on a specific address is chosen in preference over sockets
that listen on any address. The test covers UDP/UDP6/TCP/TCP6.

It is based on, and similar to, reuseport_dualstack.c selftest.

Signed-off-by: Peter Oskolkov <posk@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

4 files changed, 271 insertions, 2 deletions

diff --git a/tools/testing/selftests/net/.gitignore b/tools/testing/selftests/net/.gitignore
index 7f57b916e6b2..6f81130605d7 100644
--- a/tools/testing/selftests/net/.gitignore
+++ b/tools/testing/selftests/net/.gitignore
@@ -3,6 +3,7 @@ socket
 psock_fanout
 psock_snd
 psock_tpacket
+reuseport_addr_any
 reuseport_bpf
 reuseport_bpf_cpu
 reuseport_bpf_numa
diff --git a/tools/testing/selftests/net/Makefile b/tools/testing/selftests/net/Makefile
index ee2e27b1cd0d..aeecc3ef53d0 100644
--- a/tools/testing/selftests/net/Makefile
+++ b/tools/testing/selftests/net/Makefile
@@ -7,10 +7,10 @@ CFLAGS += -I../../../../usr/include/
 TEST_PROGS := run_netsocktests run_afpackettests test_bpf.sh netdevice.sh rtnetlink.sh
 TEST_PROGS += fib_tests.sh fib-onlink-tests.sh pmtu.sh udpgso.sh ip_defrag.sh
 TEST_PROGS += udpgso_bench.sh fib_rule_tests.sh msg_zerocopy.sh psock_snd.sh
-TEST_PROGS += udpgro_bench.sh udpgro.sh test_vxlan_under_vrf.sh
+TEST_PROGS += udpgro_bench.sh udpgro.sh test_vxlan_under_vrf.sh reuseport_addr_any.sh
 TEST_PROGS_EXTENDED := in_netns.sh
 TEST_GEN_FILES =  socket
-TEST_GEN_FILES += psock_fanout psock_tpacket msg_zerocopy
+TEST_GEN_FILES += psock_fanout psock_tpacket msg_zerocopy reuseport_addr_any
 TEST_GEN_FILES += tcp_mmap tcp_inq psock_snd txring_overwrite
 TEST_GEN_FILES += udpgso udpgso_bench_tx udpgso_bench_rx ip_defrag
 TEST_GEN_PROGS = reuseport_bpf reuseport_bpf_cpu reuseport_bpf_numa
diff --git a/tools/testing/selftests/net/reuseport_addr_any.c b/tools/testing/selftests/net/reuseport_addr_any.c
new file mode 100644
index 000000000000..f5e01d989519
--- /dev/null
+++ b/tools/testing/selftests/net/reuseport_addr_any.c
@@ -0,0 +1,264 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/* Test that sockets listening on a specific address are preferred
+ * over sockets listening on addr_any.
+ */
+
+#define _GNU_SOURCE
+
+#include <arpa/inet.h>
+#include <errno.h>
+#include <error.h>
+#include <linux/in.h>
+#include <linux/unistd.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/epoll.h>
+#include <sys/types.h>
+#include <sys/socket.h>
+#include <unistd.h>
+
+static const char *IP4_ADDR = "127.0.0.1";
+static const char *IP6_ADDR = "::1";
+static const char *IP4_MAPPED6 = "::ffff:127.0.0.1";
+
+static const int PORT = 8888;
+
+static void build_rcv_fd(int family, int proto, int *rcv_fds, int count,
+			 const char *addr_str)
+{
+	struct sockaddr_in  addr4 = {0};
+	struct sockaddr_in6 addr6 = {0};
+	struct sockaddr *addr;
+	int opt, i, sz;
+
+	memset(&addr, 0, sizeof(addr));
+
+	switch (family) {
+	case AF_INET:
+		addr4.sin_family = family;
+		if (!addr_str)
+			addr4.sin_addr.s_addr = htonl(INADDR_ANY);
+		else if (!inet_pton(family, addr_str, &addr4.sin_addr.s_addr))
+			error(1, errno, "inet_pton failed: %s", addr_str);
+		addr4.sin_port = htons(PORT);
+		sz = sizeof(addr4);
+		addr = (struct sockaddr *)&addr4;
+		break;
+	case AF_INET6:
+		addr6.sin6_family = AF_INET6;
+		if (!addr_str)
+			addr6.sin6_addr = in6addr_any;
+		else if (!inet_pton(family, addr_str, &addr6.sin6_addr))
+			error(1, errno, "inet_pton failed: %s", addr_str);
+		addr6.sin6_port = htons(PORT);
+		sz = sizeof(addr6);
+		addr = (struct sockaddr *)&addr6;
+		break;
+	default:
+		error(1, 0, "Unsupported family %d", family);
+	}
+
+	for (i = 0; i < count; ++i) {
+		rcv_fds[i] = socket(family, proto, 0);
+		if (rcv_fds[i] < 0)
+			error(1, errno, "failed to create receive socket");
+
+		opt = 1;
+		if (setsockopt(rcv_fds[i], SOL_SOCKET, SO_REUSEPORT, &opt,
+			       sizeof(opt)))
+			error(1, errno, "failed to set SO_REUSEPORT");
+
+		if (bind(rcv_fds[i], addr, sz))
+			error(1, errno, "failed to bind receive socket");
+
+		if (proto == SOCK_STREAM && listen(rcv_fds[i], 10))
+			error(1, errno, "failed to listen on receive port");
+	}
+}
+
+static int connect_and_send(int family, int proto)
+{
+	struct sockaddr_in  saddr4 = {0};
+	struct sockaddr_in  daddr4 = {0};
+	struct sockaddr_in6 saddr6 = {0};
+	struct sockaddr_in6 daddr6 = {0};
+	struct sockaddr *saddr, *daddr;
+	int fd, sz;
+
+	switch (family) {
+	case AF_INET:
+		saddr4.sin_family = AF_INET;
+		saddr4.sin_addr.s_addr = htonl(INADDR_ANY);
+		saddr4.sin_port = 0;
+
+		daddr4.sin_family = AF_INET;
+		if (!inet_pton(family, IP4_ADDR, &daddr4.sin_addr.s_addr))
+			error(1, errno, "inet_pton failed: %s", IP4_ADDR);
+		daddr4.sin_port = htons(PORT);
+
+		sz = sizeof(saddr4);
+		saddr = (struct sockaddr *)&saddr4;
+		daddr = (struct sockaddr *)&daddr4;
+	break;
+	case AF_INET6:
+		saddr6.sin6_family = AF_INET6;
+		saddr6.sin6_addr = in6addr_any;
+
+		daddr6.sin6_family = AF_INET6;
+		if (!inet_pton(family, IP6_ADDR, &daddr6.sin6_addr))
+			error(1, errno, "inet_pton failed: %s", IP6_ADDR);
+		daddr6.sin6_port = htons(PORT);
+
+		sz = sizeof(saddr6);
+		saddr = (struct sockaddr *)&saddr6;
+		daddr = (struct sockaddr *)&daddr6;
+	break;
+	default:
+		error(1, 0, "Unsupported family %d", family);
+	}
+
+	fd = socket(family, proto, 0);
+	if (fd < 0)
+		error(1, errno, "failed to create send socket");
+
+	if (bind(fd, saddr, sz))
+		error(1, errno, "failed to bind send socket");
+
+	if (connect(fd, daddr, sz))
+		error(1, errno, "failed to connect send socket");
+
+	if (send(fd, "a", 1, 0) < 0)
+		error(1, errno, "failed to send message");
+
+	return fd;
+}
+
+static int receive_once(int epfd, int proto)
+{
+	struct epoll_event ev;
+	int i, fd;
+	char buf[8];
+
+	i = epoll_wait(epfd, &ev, 1, 3);
+	if (i < 0)
+		error(1, errno, "epoll_wait failed");
+
+	if (proto == SOCK_STREAM) {
+		fd = accept(ev.data.fd, NULL, NULL);
+		if (fd < 0)
+			error(1, errno, "failed to accept");
+		i = recv(fd, buf, sizeof(buf), 0);
+		close(fd);
+	} else {
+		i = recv(ev.data.fd, buf, sizeof(buf), 0);
+	}
+
+	if (i < 0)
+		error(1, errno, "failed to recv");
+
+	return ev.data.fd;
+}
+
+static void test(int *rcv_fds, int count, int family, int proto, int fd)
+{
+	struct epoll_event ev;
+	int epfd, i, send_fd, recv_fd;
+
+	epfd = epoll_create(1);
+	if (epfd < 0)
+		error(1, errno, "failed to create epoll");
+
+	ev.events = EPOLLIN;
+	for (i = 0; i < count; ++i) {
+		ev.data.fd = rcv_fds[i];
+		if (epoll_ctl(epfd, EPOLL_CTL_ADD, rcv_fds[i], &ev))
+			error(1, errno, "failed to register sock epoll");
+	}
+
+	send_fd = connect_and_send(family, proto);
+
+	recv_fd = receive_once(epfd, proto);
+	if (recv_fd != fd)
+		error(1, 0, "received on an unexpected socket");
+
+	close(send_fd);
+	close(epfd);
+}
+
+int main(void)
+{
+	/* Below we test that a socket listening on a specific address
+	 * is always selected in preference over a socket listening
+	 * on addr_any. Bugs where this is not the case often result
+	 * in sockets created first or last to get picked. So below
+	 * we make sure that there are always addr_any sockets created
+	 * before and after a specific socket is created.
+	 */
+	int rcv_fds[10], i;
+
+	fprintf(stderr, "---- UDP IPv4 ----\n");
+	build_rcv_fd(AF_INET, SOCK_DGRAM, rcv_fds, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_DGRAM, rcv_fds + 2, 2, NULL);
+	build_rcv_fd(AF_INET, SOCK_DGRAM, rcv_fds + 4, 1, IP4_ADDR);
+	build_rcv_fd(AF_INET, SOCK_DGRAM, rcv_fds + 5, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_DGRAM, rcv_fds + 7, 2, NULL);
+	test(rcv_fds, 9, AF_INET, SOCK_DGRAM, rcv_fds[4]);
+	for (i = 0; i < 9; ++i)
+		close(rcv_fds[i]);
+
+	fprintf(stderr, "---- UDP IPv6 ----\n");
+	build_rcv_fd(AF_INET, SOCK_DGRAM, rcv_fds, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_DGRAM, rcv_fds + 2, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_DGRAM, rcv_fds + 4, 1, IP6_ADDR);
+	build_rcv_fd(AF_INET, SOCK_DGRAM, rcv_fds + 5, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_DGRAM, rcv_fds + 7, 2, NULL);
+	test(rcv_fds, 9, AF_INET6, SOCK_DGRAM, rcv_fds[4]);
+	for (i = 0; i < 9; ++i)
+		close(rcv_fds[i]);
+
+	fprintf(stderr, "---- UDP IPv4 mapped to IPv6 ----\n");
+	build_rcv_fd(AF_INET, SOCK_DGRAM, rcv_fds, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_DGRAM, rcv_fds + 2, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_DGRAM, rcv_fds + 4, 1, IP4_MAPPED6);
+	build_rcv_fd(AF_INET, SOCK_DGRAM, rcv_fds + 5, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_DGRAM, rcv_fds + 7, 2, NULL);
+	test(rcv_fds, 9, AF_INET, SOCK_DGRAM, rcv_fds[4]);
+	for (i = 0; i < 9; ++i)
+		close(rcv_fds[i]);
+
+	fprintf(stderr, "---- TCP IPv4 ----\n");
+	build_rcv_fd(AF_INET, SOCK_STREAM, rcv_fds, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_STREAM, rcv_fds + 2, 2, NULL);
+	build_rcv_fd(AF_INET, SOCK_STREAM, rcv_fds + 4, 1, IP4_ADDR);
+	build_rcv_fd(AF_INET, SOCK_STREAM, rcv_fds + 5, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_STREAM, rcv_fds + 7, 2, NULL);
+	test(rcv_fds, 9, AF_INET, SOCK_STREAM, rcv_fds[4]);
+	for (i = 0; i < 9; ++i)
+		close(rcv_fds[i]);
+
+	fprintf(stderr, "---- TCP IPv6 ----\n");
+	build_rcv_fd(AF_INET, SOCK_STREAM, rcv_fds, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_STREAM, rcv_fds + 2, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_STREAM, rcv_fds + 4, 1, IP6_ADDR);
+	build_rcv_fd(AF_INET, SOCK_STREAM, rcv_fds + 5, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_STREAM, rcv_fds + 7, 2, NULL);
+	test(rcv_fds, 9, AF_INET6, SOCK_STREAM, rcv_fds[4]);
+	for (i = 0; i < 9; ++i)
+		close(rcv_fds[i]);
+
+	fprintf(stderr, "---- TCP IPv4 mapped to IPv6 ----\n");
+	build_rcv_fd(AF_INET, SOCK_STREAM, rcv_fds, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_STREAM, rcv_fds + 2, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_STREAM, rcv_fds + 4, 1, IP4_MAPPED6);
+	build_rcv_fd(AF_INET, SOCK_STREAM, rcv_fds + 5, 2, NULL);
+	build_rcv_fd(AF_INET6, SOCK_STREAM, rcv_fds + 7, 2, NULL);
+	test(rcv_fds, 9, AF_INET, SOCK_STREAM, rcv_fds[4]);
+	for (i = 0; i < 9; ++i)
+		close(rcv_fds[i]);
+
+	fprintf(stderr, "SUCCESS\n");
+	return 0;
+}
diff --git a/tools/testing/selftests/net/reuseport_addr_any.sh b/tools/testing/selftests/net/reuseport_addr_any.sh
new file mode 100755
index 000000000000..104592f62ad4
--- /dev/null
+++ b/tools/testing/selftests/net/reuseport_addr_any.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0
+
+./in_netns.sh ./reuseport_addr_any