net: selftests: Stress reuseport listen

This patch adds a test that has 300 VIPs listening on port 443.
Each VIP:443 will have 80 listening socks by using SO_REUSEPORT.
Thus, it will have 24000 listening socks.

Before removing the port only listening_hash, all socks will be in the
same port 443 bucket and inet_reuseport_add_sock() spends much time to
walk through the bucket.  After removing the port only listening_hash
and move all usage to the port+addr lhash2, each bucket in the
ideal case has 80 sk which is much smaller than before.

Here is the test result from a qemu:
Before: listen 24000 socks took 210.210485362 (~210s)
 After: listen 24000 socks took 0.207173      (~210ms)

Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

tools/testing/selftests/net/Makefile

@@ -38,6 +38,7 @@ TEST_PROGS += srv6_end_dt6_l3vpn_test.sh
 TEST_PROGS += vrf_strict_mode_test.sh
 TEST_PROGS += arp_ndisc_evict_nocarrier.sh
 TEST_PROGS += ndisc_unsolicited_na_test.sh
+TEST_PROGS += stress_reuseport_listen.sh
 TEST_PROGS_EXTENDED := in_netns.sh setup_loopback.sh setup_veth.sh
 TEST_PROGS_EXTENDED += toeplitz_client.sh toeplitz.sh
 TEST_GEN_FILES =  socket nettest
@@ -56,6 +57,7 @@ TEST_GEN_PROGS = reuseport_bpf reuseport_bpf_cpu reuseport_bpf_numa
 TEST_GEN_PROGS += reuseport_dualstack reuseaddr_conflict tls
 TEST_GEN_FILES += toeplitz
 TEST_GEN_FILES += cmsg_sender
+TEST_GEN_FILES += stress_reuseport_listen
 TEST_PROGS += test_vxlan_vnifiltering.sh
 
 TEST_FILES := settings

tools/testing/selftests/net/stress_reuseport_listen.c

@@ -0,0 +1,105 @@
+// SPDX-License-Identifier: GPL-2.0
+
+/* Copyright (c) 2022 Meta Platforms, Inc. and affiliates. */
+
+/* Test listening on the same port 443 with multiple VIPS.
+ * Each VIP:443 will have multiple sk listening on by using
+ * SO_REUSEPORT.
+ */
+
+#include <unistd.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <error.h>
+#include <errno.h>
+#include <time.h>
+#include <arpa/inet.h>
+
+#define IP6_LADDR_START "2401:dead::1"
+#define IP6_LPORT 443
+#define NSEC_PER_SEC 1000000000L
+#define NSEC_PER_USEC 1000L
+
+static unsigned int nr_socks_per_vip;
+static unsigned int nr_vips;
+
+static int *bind_reuseport_sock6(void)
+{
+	int *lfds, *cur_fd, err, optvalue = 1;
+	struct sockaddr_in6 sa6 = {};
+	unsigned int i, j;
+
+	sa6.sin6_family = AF_INET6;
+	sa6.sin6_port = htons(IP6_LPORT);
+	err = inet_pton(AF_INET6, IP6_LADDR_START, &sa6.sin6_addr);
+	if (err != 1)
+		error(1, err, "inet_pton(%s)", IP6_LADDR_START);
+
+	lfds = malloc(nr_vips * nr_socks_per_vip * sizeof(lfds[0]));
+	if (!lfds)
+		error(1, errno, "cannot alloc array of lfds");
+
+	cur_fd = lfds;
+	for (i = 0; i < nr_vips; i++) {
+		for (j = 0; j < nr_socks_per_vip; j++) {
+			*cur_fd = socket(AF_INET6, SOCK_STREAM, 0);
+			if (*cur_fd == -1)
+				error(1, errno,
+				      "lfds[%u,%u] = socket(AF_INET6)", i, j);
+
+			err = setsockopt(*cur_fd, SOL_SOCKET, SO_REUSEPORT,
+					 &optvalue, sizeof(optvalue));
+			if (err)
+				error(1, errno,
+				      "setsockopt(lfds[%u,%u], SO_REUSEPORT)",
+				      i, j);
+
+			err = bind(*cur_fd, (struct sockaddr *)&sa6,
+				   sizeof(sa6));
+			if (err)
+				error(1, errno, "bind(lfds[%u,%u])", i, j);
+			cur_fd++;
+		}
+		sa6.sin6_addr.s6_addr32[3]++;
+	}
+
+	return lfds;
+}
+
+int main(int argc, const char *argv[])
+{
+	struct timespec start_ts, end_ts;
+	unsigned long start_ns, end_ns;
+	unsigned int nr_lsocks;
+	int *lfds, i, err;
+
+	if (argc != 3 || atoi(argv[1]) <= 0 || atoi(argv[2]) <= 0)
+		error(1, 0, "Usage: %s <nr_vips> <nr_socks_per_vip>\n",
+		      argv[0]);
+
+	nr_vips = atoi(argv[1]);
+	nr_socks_per_vip = atoi(argv[2]);
+	nr_lsocks = nr_vips * nr_socks_per_vip;
+	lfds = bind_reuseport_sock6();
+
+	clock_gettime(CLOCK_MONOTONIC, &start_ts);
+	for (i = 0; i < nr_lsocks; i++) {
+		err = listen(lfds[i], 0);
+		if (err)
+			error(1, errno, "listen(lfds[%d])", i);
+	}
+	clock_gettime(CLOCK_MONOTONIC, &end_ts);
+
+	start_ns = start_ts.tv_sec * NSEC_PER_SEC + start_ts.tv_nsec;
+	end_ns = end_ts.tv_sec * NSEC_PER_SEC + end_ts.tv_nsec;
+
+	printf("listen %d socks took %lu.%lu\n", nr_lsocks,
+	       (end_ns - start_ns) / NSEC_PER_SEC,
+	       (end_ns - start_ns) / NSEC_PER_USEC);
+
+	for (i = 0; i < nr_lsocks; i++)
+		close(lfds[i]);
+
+	free(lfds);
+	return 0;
+}

tools/testing/selftests/net/stress_reuseport_listen.sh

@@ -0,0 +1,25 @@
+#!/bin/bash
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (c) 2022 Meta Platforms, Inc. and affiliates.
+
+NS='stress_reuseport_listen_ns'
+NR_FILES=24100
+SAVED_NR_FILES=$(ulimit -n)
+
+setup() {
+	ip netns add $NS
+	ip netns exec $NS sysctl -q -w net.ipv6.ip_nonlocal_bind=1
+	ulimit -n $NR_FILES
+}
+
+cleanup() {
+	ip netns del $NS
+	ulimit -n $SAVED_NR_FILES
+}
+
+trap cleanup EXIT
+setup
+# 300 different vips listen on port 443
+# Each vip:443 sockaddr has 80 LISTEN sock by using SO_REUSEPORT
+# Total 24000 listening socks
+ip netns exec $NS ./stress_reuseport_listen 300 80