This patch adds tests to exercise optnames that are allowed in bpf_setsockopt(). Reviewed-by: Stanislav Fomichev <sdf@xxxxxxxxxx> Signed-off-by: Martin KaFai Lau <kafai@xxxxxx> --- .../selftests/bpf/prog_tests/setget_sockopt.c | 125 +++++ .../selftests/bpf/progs/bpf_tracing_net.h | 31 +- .../selftests/bpf/progs/setget_sockopt.c | 451 ++++++++++++++++++ 3 files changed, 606 insertions(+), 1 deletion(-) create mode 100644 tools/testing/selftests/bpf/prog_tests/setget_sockopt.c create mode 100644 tools/testing/selftests/bpf/progs/setget_sockopt.c diff --git a/tools/testing/selftests/bpf/prog_tests/setget_sockopt.c b/tools/testing/selftests/bpf/prog_tests/setget_sockopt.c new file mode 100644 index 000000000000..018611e6b248 --- /dev/null +++ b/tools/testing/selftests/bpf/prog_tests/setget_sockopt.c @@ -0,0 +1,125 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (c) Meta Platforms, Inc. and affiliates. */ + +#define _GNU_SOURCE +#include <sched.h> +#include <linux/socket.h> +#include <net/if.h> + +#include "test_progs.h" +#include "cgroup_helpers.h" +#include "network_helpers.h" + +#include "setget_sockopt.skel.h" + +#define CG_NAME "/setget-sockopt-test" + +static const char addr4_str[] = "127.0.0.1"; +static const char addr6_str[] = "::1"; +static struct setget_sockopt *skel; +static int cg_fd; + +static int create_netns(void) +{ + if (!ASSERT_OK(unshare(CLONE_NEWNET), "create netns")) + return -1; + + if (!ASSERT_OK(system("ip link set dev lo up"), "set lo up")) + return -1; + + if (!ASSERT_OK(system("ip link add dev binddevtest1 type veth peer name binddevtest2"), + "add veth")) + return -1; + + if (!ASSERT_OK(system("ip link set dev binddevtest1 up"), + "bring veth up")) + return -1; + + return 0; +} + +static void test_tcp(int family) +{ + struct setget_sockopt__bss *bss = skel->bss; + int sfd, cfd; + + memset(bss, 0, sizeof(*bss)); + + sfd = start_server(family, SOCK_STREAM, + family == AF_INET6 ? addr6_str : addr4_str, 0, 0); + if (!ASSERT_GE(sfd, 0, "start_server")) + return; + + cfd = connect_to_fd(sfd, 0); + if (!ASSERT_GE(cfd, 0, "connect_to_fd_server")) { + close(sfd); + return; + } + close(sfd); + close(cfd); + + ASSERT_EQ(bss->nr_listen, 1, "nr_listen"); + ASSERT_EQ(bss->nr_connect, 1, "nr_connect"); + ASSERT_EQ(bss->nr_active, 1, "nr_active"); + ASSERT_EQ(bss->nr_passive, 1, "nr_passive"); + ASSERT_EQ(bss->nr_socket_post_create, 2, "nr_socket_post_create"); + ASSERT_EQ(bss->nr_binddev, 2, "nr_bind"); +} + +static void test_udp(int family) +{ + struct setget_sockopt__bss *bss = skel->bss; + int sfd; + + memset(bss, 0, sizeof(*bss)); + + sfd = start_server(family, SOCK_DGRAM, + family == AF_INET6 ? addr6_str : addr4_str, 0, 0); + if (!ASSERT_GE(sfd, 0, "start_server")) + return; + close(sfd); + + ASSERT_GE(bss->nr_socket_post_create, 1, "nr_socket_post_create"); + ASSERT_EQ(bss->nr_binddev, 1, "nr_bind"); +} + +void test_setget_sockopt(void) +{ + cg_fd = test__join_cgroup(CG_NAME); + if (cg_fd < 0) + return; + + if (create_netns()) + goto done; + + skel = setget_sockopt__open(); + if (!ASSERT_OK_PTR(skel, "open skel")) + goto done; + + strcpy(skel->rodata->veth, "binddevtest1"); + skel->rodata->veth_ifindex = if_nametoindex("binddevtest1"); + if (!ASSERT_GT(skel->rodata->veth_ifindex, 0, "if_nametoindex")) + goto done; + + if (!ASSERT_OK(setget_sockopt__load(skel), "load skel")) + goto done; + + skel->links.skops_sockopt = + bpf_program__attach_cgroup(skel->progs.skops_sockopt, cg_fd); + if (!ASSERT_OK_PTR(skel->links.skops_sockopt, "attach cgroup")) + goto done; + + skel->links.socket_post_create = + bpf_program__attach_cgroup(skel->progs.socket_post_create, cg_fd); + if (!ASSERT_OK_PTR(skel->links.socket_post_create, "attach_cgroup")) + goto done; + + test_tcp(AF_INET6); + test_tcp(AF_INET); + test_udp(AF_INET6); + test_udp(AF_INET); + +done: + setget_sockopt__destroy(skel); + close(cg_fd); +} diff --git a/tools/testing/selftests/bpf/progs/bpf_tracing_net.h b/tools/testing/selftests/bpf/progs/bpf_tracing_net.h index 98dd2c4815f0..5ebc6dabef84 100644 --- a/tools/testing/selftests/bpf/progs/bpf_tracing_net.h +++ b/tools/testing/selftests/bpf/progs/bpf_tracing_net.h @@ -6,13 +6,40 @@ #define AF_INET6 10 #define SOL_SOCKET 1 +#define SO_REUSEADDR 2 #define SO_SNDBUF 7 -#define __SO_ACCEPTCON (1 << 16) +#define SO_RCVBUF 8 +#define SO_KEEPALIVE 9 #define SO_PRIORITY 12 +#define SO_REUSEPORT 15 +#define SO_RCVLOWAT 18 +#define SO_BINDTODEVICE 25 +#define SO_MARK 36 +#define SO_MAX_PACING_RATE 47 +#define SO_BINDTOIFINDEX 62 +#define SO_TXREHASH 74 +#define __SO_ACCEPTCON (1 << 16) + +#define IP_TOS 1 + +#define IPV6_TCLASS 67 +#define IPV6_AUTOFLOWLABEL 70 #define SOL_TCP 6 +#define TCP_NODELAY 1 +#define TCP_MAXSEG 2 +#define TCP_KEEPIDLE 4 +#define TCP_KEEPINTVL 5 +#define TCP_KEEPCNT 6 +#define TCP_SYNCNT 7 +#define TCP_WINDOW_CLAMP 10 #define TCP_CONGESTION 13 +#define TCP_THIN_LINEAR_TIMEOUTS 16 +#define TCP_USER_TIMEOUT 18 +#define TCP_NOTSENT_LOWAT 25 +#define TCP_SAVE_SYN 27 #define TCP_CA_NAME_MAX 16 +#define TCP_NAGLE_OFF 1 #define ICSK_TIME_RETRANS 1 #define ICSK_TIME_PROBE0 3 @@ -49,6 +76,8 @@ #define sk_state __sk_common.skc_state #define sk_v6_daddr __sk_common.skc_v6_daddr #define sk_v6_rcv_saddr __sk_common.skc_v6_rcv_saddr +#define sk_flags __sk_common.skc_flags +#define sk_reuse __sk_common.skc_reuse #define s6_addr32 in6_u.u6_addr32 diff --git a/tools/testing/selftests/bpf/progs/setget_sockopt.c b/tools/testing/selftests/bpf/progs/setget_sockopt.c new file mode 100644 index 000000000000..4a4cb44a4a15 --- /dev/null +++ b/tools/testing/selftests/bpf/progs/setget_sockopt.c @@ -0,0 +1,451 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (c) Meta Platforms, Inc. and affiliates. */ + +#include "vmlinux.h" +#include "bpf_tracing_net.h" +#include <bpf/bpf_core_read.h> +#include <bpf/bpf_helpers.h> +#include <bpf/bpf_tracing.h> + +#ifndef ARRAY_SIZE +#define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) +#endif + +extern unsigned long CONFIG_HZ __kconfig; + +const volatile char veth[IFNAMSIZ]; +const volatile int veth_ifindex; + +int nr_listen; +int nr_passive; +int nr_active; +int nr_connect; +int nr_binddev; +int nr_socket_post_create; + +struct sockopt_test { + int opt; + int new; + int restore; + int expected; + int tcp_expected; + unsigned int flip:1; +}; + +static const char cubic_cc[] = "cubic"; +static const char reno_cc[] = "reno"; + +static const struct sockopt_test sol_socket_tests[] = { + { .opt = SO_REUSEADDR, .flip = 1, }, + { .opt = SO_SNDBUF, .new = 8123, .expected = 8123 * 2, }, + { .opt = SO_RCVBUF, .new = 8123, .expected = 8123 * 2, }, + { .opt = SO_KEEPALIVE, .flip = 1, }, + { .opt = SO_PRIORITY, .new = 0xeb9f, .expected = 0xeb9f, }, + { .opt = SO_REUSEPORT, .flip = 1, }, + { .opt = SO_RCVLOWAT, .new = 8123, .expected = 8123, }, + { .opt = SO_MARK, .new = 0xeb9f, .expected = 0xeb9f, }, + { .opt = SO_MAX_PACING_RATE, .new = 0xeb9f, .expected = 0xeb9f, }, + { .opt = SO_TXREHASH, .flip = 1, }, + { .opt = 0, }, +}; + +static const struct sockopt_test sol_tcp_tests[] = { + { .opt = TCP_NODELAY, .flip = 1, }, + { .opt = TCP_MAXSEG, .new = 1314, .expected = 1314, }, + { .opt = TCP_KEEPIDLE, .new = 123, .expected = 123, .restore = 321, }, + { .opt = TCP_KEEPINTVL, .new = 123, .expected = 123, .restore = 321, }, + { .opt = TCP_KEEPCNT, .new = 123, .expected = 123, .restore = 124, }, + { .opt = TCP_SYNCNT, .new = 123, .expected = 123, .restore = 124, }, + { .opt = TCP_WINDOW_CLAMP, .new = 8123, .expected = 8123, .restore = 8124, }, + { .opt = TCP_CONGESTION, }, + { .opt = TCP_THIN_LINEAR_TIMEOUTS, .flip = 1, }, + { .opt = TCP_USER_TIMEOUT, .new = 123400, .expected = 123400, }, + { .opt = TCP_NOTSENT_LOWAT, .new = 1314, .expected = 1314, }, + { .opt = TCP_SAVE_SYN, .new = 1, .expected = 1, }, + { .opt = 0, }, +}; + +static const struct sockopt_test sol_ip_tests[] = { + { .opt = IP_TOS, .new = 0xe1, .expected = 0xe1, .tcp_expected = 0xe0, }, + { .opt = 0, }, +}; + +static const struct sockopt_test sol_ipv6_tests[] = { + { .opt = IPV6_TCLASS, .new = 0xe1, .expected = 0xe1, .tcp_expected = 0xe0, }, + { .opt = IPV6_AUTOFLOWLABEL, .flip = 1, }, + { .opt = 0, }, +}; + +struct loop_ctx { + void *ctx; + struct sock *sk; +}; + +static int __bpf_getsockopt(void *ctx, struct sock *sk, + int level, int opt, int *optval, + int optlen) +{ + if (level == SOL_SOCKET) { + switch (opt) { + case SO_REUSEADDR: + *optval = !!BPF_CORE_READ_BITFIELD(sk, sk_reuse); + break; + case SO_KEEPALIVE: + *optval = !!(sk->sk_flags & (1UL << 3)); + break; + case SO_RCVLOWAT: + *optval = sk->sk_rcvlowat; + break; + case SO_MAX_PACING_RATE: + *optval = sk->sk_max_pacing_rate; + break; + default: + return bpf_getsockopt(ctx, level, opt, optval, optlen); + } + return 0; + } + + if (level == IPPROTO_TCP) { + struct tcp_sock *tp = bpf_skc_to_tcp_sock(sk); + + if (!tp) + return -1; + + switch (opt) { + case TCP_NODELAY: + *optval = !!(BPF_CORE_READ_BITFIELD(tp, nonagle) & TCP_NAGLE_OFF); + break; + case TCP_MAXSEG: + *optval = tp->rx_opt.user_mss; + break; + case TCP_KEEPIDLE: + *optval = tp->keepalive_time / CONFIG_HZ; + break; + case TCP_SYNCNT: + *optval = tp->inet_conn.icsk_syn_retries; + break; + case TCP_KEEPINTVL: + *optval = tp->keepalive_intvl / CONFIG_HZ; + break; + case TCP_KEEPCNT: + *optval = tp->keepalive_probes; + break; + case TCP_WINDOW_CLAMP: + *optval = tp->window_clamp; + break; + case TCP_THIN_LINEAR_TIMEOUTS: + *optval = !!BPF_CORE_READ_BITFIELD(tp, thin_lto); + break; + case TCP_USER_TIMEOUT: + *optval = tp->inet_conn.icsk_user_timeout; + break; + case TCP_NOTSENT_LOWAT: + *optval = tp->notsent_lowat; + break; + case TCP_SAVE_SYN: + *optval = BPF_CORE_READ_BITFIELD(tp, save_syn); + break; + default: + return bpf_getsockopt(ctx, level, opt, optval, optlen); + } + return 0; + } + + if (level == IPPROTO_IPV6) { + switch (opt) { + case IPV6_AUTOFLOWLABEL: { + __u16 proto = sk->sk_protocol; + struct inet_sock *inet_sk; + + if (proto == IPPROTO_TCP) + inet_sk = (struct inet_sock *)bpf_skc_to_tcp_sock(sk); + else + inet_sk = (struct inet_sock *)bpf_skc_to_udp6_sock(sk); + + if (!inet_sk) + return -1; + + *optval = !!inet_sk->pinet6->autoflowlabel; + break; + } + default: + return bpf_getsockopt(ctx, level, opt, optval, optlen); + } + return 0; + } + + return bpf_getsockopt(ctx, level, opt, optval, optlen); +} + +static int bpf_test_sockopt_flip(void *ctx, struct sock *sk, + const struct sockopt_test *t, + int level) +{ + int old, tmp, new, opt = t->opt; + + opt = t->opt; + + if (__bpf_getsockopt(ctx, sk, level, opt, &old, sizeof(old))) + return 1; + /* kernel initialized txrehash to 255 */ + if (level == SOL_SOCKET && opt == SO_TXREHASH && old != 0 && old != 1) + old = 1; + + new = !old; + if (bpf_setsockopt(ctx, level, opt, &new, sizeof(new))) + return 1; + if (__bpf_getsockopt(ctx, sk, level, opt, &tmp, sizeof(tmp)) || + tmp != new) + return 1; + + if (bpf_setsockopt(ctx, level, opt, &old, sizeof(old))) + return 1; + + return 0; +} + +static int bpf_test_sockopt_int(void *ctx, struct sock *sk, + const struct sockopt_test *t, + int level) +{ + int old, tmp, new, expected, opt; + + opt = t->opt; + new = t->new; + if (sk->sk_type == SOCK_STREAM && t->tcp_expected) + expected = t->tcp_expected; + else + expected = t->expected; + + if (__bpf_getsockopt(ctx, sk, level, opt, &old, sizeof(old)) || + old == new) + return 1; + + if (bpf_setsockopt(ctx, level, opt, &new, sizeof(new))) + return 1; + if (__bpf_getsockopt(ctx, sk, level, opt, &tmp, sizeof(tmp)) || + tmp != expected) + return 1; + + if (t->restore) + old = t->restore; + if (bpf_setsockopt(ctx, level, opt, &old, sizeof(old))) + return 1; + + return 0; +} + +static int bpf_test_socket_sockopt(__u32 i, struct loop_ctx *lc) +{ + const struct sockopt_test *t; + + if (i >= ARRAY_SIZE(sol_socket_tests)) + return 1; + + t = &sol_socket_tests[i]; + if (!t->opt) + return 1; + + if (t->flip) + return bpf_test_sockopt_flip(lc->ctx, lc->sk, t, SOL_SOCKET); + + return bpf_test_sockopt_int(lc->ctx, lc->sk, t, SOL_SOCKET); +} + +static int bpf_test_ip_sockopt(__u32 i, struct loop_ctx *lc) +{ + const struct sockopt_test *t; + + if (i >= ARRAY_SIZE(sol_ip_tests)) + return 1; + + t = &sol_ip_tests[i]; + if (!t->opt) + return 1; + + if (t->flip) + return bpf_test_sockopt_flip(lc->ctx, lc->sk, t, IPPROTO_IP); + + return bpf_test_sockopt_int(lc->ctx, lc->sk, t, IPPROTO_IP); +} + +static int bpf_test_ipv6_sockopt(__u32 i, struct loop_ctx *lc) +{ + const struct sockopt_test *t; + + if (i >= ARRAY_SIZE(sol_ipv6_tests)) + return 1; + + t = &sol_ipv6_tests[i]; + if (!t->opt) + return 1; + + if (t->flip) + return bpf_test_sockopt_flip(lc->ctx, lc->sk, t, IPPROTO_IPV6); + + return bpf_test_sockopt_int(lc->ctx, lc->sk, t, IPPROTO_IPV6); +} + +static int bpf_test_tcp_sockopt(__u32 i, struct loop_ctx *lc) +{ + const struct sockopt_test *t; + struct sock *sk; + void *ctx; + + if (i >= ARRAY_SIZE(sol_tcp_tests)) + return 1; + + t = &sol_tcp_tests[i]; + if (!t->opt) + return 1; + + ctx = lc->ctx; + sk = lc->sk; + + if (t->opt == TCP_CONGESTION) { + char old_cc[16], tmp_cc[16]; + const char *new_cc; + + if (bpf_getsockopt(ctx, IPPROTO_TCP, TCP_CONGESTION, old_cc, sizeof(old_cc))) + return 1; + if (!bpf_strncmp(old_cc, sizeof(old_cc), cubic_cc)) + new_cc = reno_cc; + else + new_cc = cubic_cc; + if (bpf_setsockopt(ctx, IPPROTO_TCP, TCP_CONGESTION, (void *)new_cc, + sizeof(new_cc))) + return 1; + if (bpf_getsockopt(ctx, IPPROTO_TCP, TCP_CONGESTION, tmp_cc, sizeof(tmp_cc))) + return 1; + if (bpf_strncmp(tmp_cc, sizeof(tmp_cc), new_cc)) + return 1; + if (bpf_setsockopt(ctx, IPPROTO_TCP, TCP_CONGESTION, old_cc, sizeof(old_cc))) + return 1; + return 0; + } + + if (t->flip) + return bpf_test_sockopt_flip(ctx, sk, t, IPPROTO_TCP); + + return bpf_test_sockopt_int(ctx, sk, t, IPPROTO_TCP); +} + +static int bpf_test_sockopt(void *ctx, struct sock *sk) +{ + struct loop_ctx lc = { .ctx = ctx, .sk = sk, }; + __u16 family, proto; + int n; + + family = sk->sk_family; + proto = sk->sk_protocol; + + n = bpf_loop(ARRAY_SIZE(sol_socket_tests), bpf_test_socket_sockopt, &lc, 0); + if (n != ARRAY_SIZE(sol_socket_tests)) + return -1; + + if (proto == IPPROTO_TCP) { + n = bpf_loop(ARRAY_SIZE(sol_tcp_tests), bpf_test_tcp_sockopt, &lc, 0); + if (n != ARRAY_SIZE(sol_tcp_tests)) + return -1; + } + + if (family == AF_INET) { + n = bpf_loop(ARRAY_SIZE(sol_ip_tests), bpf_test_ip_sockopt, &lc, 0); + if (n != ARRAY_SIZE(sol_ip_tests)) + return -1; + } else { + n = bpf_loop(ARRAY_SIZE(sol_ipv6_tests), bpf_test_ipv6_sockopt, &lc, 0); + if (n != ARRAY_SIZE(sol_ipv6_tests)) + return -1; + } + + return 0; +} + +static int binddev_test(void *ctx) +{ + const char empty_ifname[] = ""; + int ifindex, zero = 0; + + if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE, + (void *)veth, sizeof(veth))) + return -1; + if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX, + &ifindex, sizeof(int)) || + ifindex != veth_ifindex) + return -1; + + if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTODEVICE, + (void *)empty_ifname, sizeof(empty_ifname))) + return -1; + if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX, + &ifindex, sizeof(int)) || + ifindex != 0) + return -1; + + if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX, + (void *)&veth_ifindex, sizeof(int))) + return -1; + if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX, + &ifindex, sizeof(int)) || + ifindex != veth_ifindex) + return -1; + + if (bpf_setsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX, + &zero, sizeof(int))) + return -1; + if (bpf_getsockopt(ctx, SOL_SOCKET, SO_BINDTOIFINDEX, + &ifindex, sizeof(int)) || + ifindex != 0) + return -1; + + return 0; +} + +SEC("lsm_cgroup/socket_post_create") +int BPF_PROG(socket_post_create, struct socket *sock, int family, + int type, int protocol, int kern) +{ + struct sock *sk = sock->sk; + + if (!sk) + return 1; + + nr_socket_post_create += !bpf_test_sockopt(sk, sk); + nr_binddev += !binddev_test(sk); + + return 1; +} + +SEC("sockops") +int skops_sockopt(struct bpf_sock_ops *skops) +{ + struct bpf_sock *bpf_sk = skops->sk; + struct sock *sk; + + if (!bpf_sk) + return 1; + + sk = (struct sock *)bpf_skc_to_tcp_sock(bpf_sk); + if (!sk) + return 1; + + switch (skops->op) { + case BPF_SOCK_OPS_TCP_LISTEN_CB: + nr_listen += !bpf_test_sockopt(skops, sk); + break; + case BPF_SOCK_OPS_TCP_CONNECT_CB: + nr_connect += !bpf_test_sockopt(skops, sk); + break; + case BPF_SOCK_OPS_ACTIVE_ESTABLISHED_CB: + nr_active += !bpf_test_sockopt(skops, sk); + break; + case BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB: + nr_passive += !bpf_test_sockopt(skops, sk); + break; + } + + return 1; +} + +char _license[] SEC("license") = "GPL"; -- 2.30.2