From e0b27b3f97b8fce620331baad563833617c1f303 Mon Sep 17 00:00:00 2001 From: Martin KaFai Lau Date: Sat, 9 Feb 2019 23:22:28 -0800 Subject: bpf: Add test_sock_fields for skb->sk and bpf_tcp_sock This patch adds a C program to show the usage on skb->sk and bpf_tcp_sock. Acked-by: Alexei Starovoitov Signed-off-by: Martin KaFai Lau Signed-off-by: Alexei Starovoitov --- tools/testing/selftests/bpf/Makefile | 6 +- tools/testing/selftests/bpf/bpf_helpers.h | 4 + tools/testing/selftests/bpf/test_sock_fields.c | 327 +++++++++++++++++++++ .../testing/selftests/bpf/test_sock_fields_kern.c | 152 ++++++++++ 4 files changed, 487 insertions(+), 2 deletions(-) create mode 100644 tools/testing/selftests/bpf/test_sock_fields.c create mode 100644 tools/testing/selftests/bpf/test_sock_fields_kern.c (limited to 'tools') diff --git a/tools/testing/selftests/bpf/Makefile b/tools/testing/selftests/bpf/Makefile index 383d2ff13fc7..c7e1e3255448 100644 --- a/tools/testing/selftests/bpf/Makefile +++ b/tools/testing/selftests/bpf/Makefile @@ -23,7 +23,7 @@ TEST_GEN_PROGS = test_verifier test_tag test_maps test_lru_map test_lpm_map test test_align test_verifier_log test_dev_cgroup test_tcpbpf_user \ test_sock test_btf test_sockmap test_lirc_mode2_user get_cgroup_id_user \ test_socket_cookie test_cgroup_storage test_select_reuseport test_section_names \ - test_netcnt test_tcpnotify_user + test_netcnt test_tcpnotify_user test_sock_fields BPF_OBJ_FILES = \ test_xdp_redirect.o test_xdp_meta.o sockmap_parse_prog.o \ @@ -35,7 +35,8 @@ BPF_OBJ_FILES = \ sendmsg4_prog.o sendmsg6_prog.o test_lirc_mode2_kern.o \ get_cgroup_id_kern.o socket_cookie_prog.o test_select_reuseport_kern.o \ test_skb_cgroup_id_kern.o bpf_flow.o netcnt_prog.o test_xdp_vlan.o \ - xdp_dummy.o test_map_in_map.o test_spin_lock.o test_map_lock.o + xdp_dummy.o test_map_in_map.o test_spin_lock.o test_map_lock.o \ + test_sock_fields_kern.o # Objects are built with default compilation flags and with sub-register # code-gen enabled. @@ -111,6 +112,7 @@ $(OUTPUT)/test_progs: trace_helpers.c $(OUTPUT)/get_cgroup_id_user: cgroup_helpers.c $(OUTPUT)/test_cgroup_storage: cgroup_helpers.c $(OUTPUT)/test_netcnt: cgroup_helpers.c +$(OUTPUT)/test_sock_fields: cgroup_helpers.c .PHONY: force diff --git a/tools/testing/selftests/bpf/bpf_helpers.h b/tools/testing/selftests/bpf/bpf_helpers.h index 6a0ce0f055c5..d9999f1ed1d2 100644 --- a/tools/testing/selftests/bpf/bpf_helpers.h +++ b/tools/testing/selftests/bpf/bpf_helpers.h @@ -176,6 +176,10 @@ static void (*bpf_spin_lock)(struct bpf_spin_lock *lock) = (void *) BPF_FUNC_spin_lock; static void (*bpf_spin_unlock)(struct bpf_spin_lock *lock) = (void *) BPF_FUNC_spin_unlock; +static struct bpf_sock *(*bpf_sk_fullsock)(struct bpf_sock *sk) = + (void *) BPF_FUNC_sk_fullsock; +static struct bpf_tcp_sock *(*bpf_tcp_sock)(struct bpf_sock *sk) = + (void *) BPF_FUNC_tcp_sock; /* llvm builtin functions that eBPF C program may use to * emit BPF_LD_ABS and BPF_LD_IND instructions diff --git a/tools/testing/selftests/bpf/test_sock_fields.c b/tools/testing/selftests/bpf/test_sock_fields.c new file mode 100644 index 000000000000..9bb58369b481 --- /dev/null +++ b/tools/testing/selftests/bpf/test_sock_fields.c @@ -0,0 +1,327 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (c) 2019 Facebook */ + +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +#include "cgroup_helpers.h" + +enum bpf_array_idx { + SRV_IDX, + CLI_IDX, + __NR_BPF_ARRAY_IDX, +}; + +#define CHECK(condition, tag, format...) ({ \ + int __ret = !!(condition); \ + if (__ret) { \ + printf("%s(%d):FAIL:%s ", __func__, __LINE__, tag); \ + printf(format); \ + printf("\n"); \ + exit(-1); \ + } \ +}) + +#define TEST_CGROUP "/test-bpf-sock-fields" +#define DATA "Hello BPF!" +#define DATA_LEN sizeof(DATA) + +static struct sockaddr_in6 srv_sa6, cli_sa6; +static int linum_map_fd; +static int addr_map_fd; +static int tp_map_fd; +static int sk_map_fd; +static __u32 srv_idx = SRV_IDX; +static __u32 cli_idx = CLI_IDX; + +static void init_loopback6(struct sockaddr_in6 *sa6) +{ + memset(sa6, 0, sizeof(*sa6)); + sa6->sin6_family = AF_INET6; + sa6->sin6_addr = in6addr_loopback; +} + +static void print_sk(const struct bpf_sock *sk) +{ + char src_ip4[24], dst_ip4[24]; + char src_ip6[64], dst_ip6[64]; + + inet_ntop(AF_INET, &sk->src_ip4, src_ip4, sizeof(src_ip4)); + inet_ntop(AF_INET6, &sk->src_ip6, src_ip6, sizeof(src_ip6)); + inet_ntop(AF_INET, &sk->dst_ip4, dst_ip4, sizeof(dst_ip4)); + inet_ntop(AF_INET6, &sk->dst_ip6, dst_ip6, sizeof(dst_ip6)); + + printf("state:%u bound_dev_if:%u family:%u type:%u protocol:%u mark:%u priority:%u " + "src_ip4:%x(%s) src_ip6:%x:%x:%x:%x(%s) src_port:%u " + "dst_ip4:%x(%s) dst_ip6:%x:%x:%x:%x(%s) dst_port:%u\n", + sk->state, sk->bound_dev_if, sk->family, sk->type, sk->protocol, + sk->mark, sk->priority, + sk->src_ip4, src_ip4, + sk->src_ip6[0], sk->src_ip6[1], sk->src_ip6[2], sk->src_ip6[3], + src_ip6, sk->src_port, + sk->dst_ip4, dst_ip4, + sk->dst_ip6[0], sk->dst_ip6[1], sk->dst_ip6[2], sk->dst_ip6[3], + dst_ip6, ntohs(sk->dst_port)); +} + +static void print_tp(const struct bpf_tcp_sock *tp) +{ + printf("snd_cwnd:%u srtt_us:%u rtt_min:%u snd_ssthresh:%u rcv_nxt:%u " + "snd_nxt:%u snd:una:%u mss_cache:%u ecn_flags:%u " + "rate_delivered:%u rate_interval_us:%u packets_out:%u " + "retrans_out:%u total_retrans:%u segs_in:%u data_segs_in:%u " + "segs_out:%u data_segs_out:%u lost_out:%u sacked_out:%u " + "bytes_received:%llu bytes_acked:%llu\n", + tp->snd_cwnd, tp->srtt_us, tp->rtt_min, tp->snd_ssthresh, + tp->rcv_nxt, tp->snd_nxt, tp->snd_una, tp->mss_cache, + tp->ecn_flags, tp->rate_delivered, tp->rate_interval_us, + tp->packets_out, tp->retrans_out, tp->total_retrans, + tp->segs_in, tp->data_segs_in, tp->segs_out, + tp->data_segs_out, tp->lost_out, tp->sacked_out, + tp->bytes_received, tp->bytes_acked); +} + +static void check_result(void) +{ + struct bpf_tcp_sock srv_tp, cli_tp; + struct bpf_sock srv_sk, cli_sk; + __u32 linum, idx0 = 0; + int err; + + err = bpf_map_lookup_elem(linum_map_fd, &idx0, &linum); + CHECK(err == -1, "bpf_map_lookup_elem(linum_map_fd)", + "err:%d errno:%d", err, errno); + + err = bpf_map_lookup_elem(sk_map_fd, &srv_idx, &srv_sk); + CHECK(err == -1, "bpf_map_lookup_elem(sk_map_fd, &srv_idx)", + "err:%d errno:%d", err, errno); + err = bpf_map_lookup_elem(tp_map_fd, &srv_idx, &srv_tp); + CHECK(err == -1, "bpf_map_lookup_elem(tp_map_fd, &srv_idx)", + "err:%d errno:%d", err, errno); + + err = bpf_map_lookup_elem(sk_map_fd, &cli_idx, &cli_sk); + CHECK(err == -1, "bpf_map_lookup_elem(sk_map_fd, &cli_idx)", + "err:%d errno:%d", err, errno); + err = bpf_map_lookup_elem(tp_map_fd, &cli_idx, &cli_tp); + CHECK(err == -1, "bpf_map_lookup_elem(tp_map_fd, &cli_idx)", + "err:%d errno:%d", err, errno); + + printf("srv_sk: "); + print_sk(&srv_sk); + printf("\n"); + + printf("cli_sk: "); + print_sk(&cli_sk); + printf("\n"); + + printf("srv_tp: "); + print_tp(&srv_tp); + printf("\n"); + + printf("cli_tp: "); + print_tp(&cli_tp); + printf("\n"); + + CHECK(srv_sk.state == 10 || + !srv_sk.state || + srv_sk.family != AF_INET6 || + srv_sk.protocol != IPPROTO_TCP || + memcmp(srv_sk.src_ip6, &in6addr_loopback, + sizeof(srv_sk.src_ip6)) || + memcmp(srv_sk.dst_ip6, &in6addr_loopback, + sizeof(srv_sk.dst_ip6)) || + srv_sk.src_port != ntohs(srv_sa6.sin6_port) || + srv_sk.dst_port != cli_sa6.sin6_port, + "Unexpected srv_sk", "Check srv_sk output. linum:%u", linum); + + CHECK(cli_sk.state == 10 || + !cli_sk.state || + cli_sk.family != AF_INET6 || + cli_sk.protocol != IPPROTO_TCP || + memcmp(cli_sk.src_ip6, &in6addr_loopback, + sizeof(cli_sk.src_ip6)) || + memcmp(cli_sk.dst_ip6, &in6addr_loopback, + sizeof(cli_sk.dst_ip6)) || + cli_sk.src_port != ntohs(cli_sa6.sin6_port) || + cli_sk.dst_port != srv_sa6.sin6_port, + "Unexpected cli_sk", "Check cli_sk output. linum:%u", linum); + + CHECK(srv_tp.data_segs_out != 1 || + srv_tp.data_segs_in || + srv_tp.snd_cwnd != 10 || + srv_tp.total_retrans || + srv_tp.bytes_acked != DATA_LEN, + "Unexpected srv_tp", "Check srv_tp output. linum:%u", linum); + + CHECK(cli_tp.data_segs_out || + cli_tp.data_segs_in != 1 || + cli_tp.snd_cwnd != 10 || + cli_tp.total_retrans || + cli_tp.bytes_received != DATA_LEN, + "Unexpected cli_tp", "Check cli_tp output. linum:%u", linum); +} + +static void test(void) +{ + int listen_fd, cli_fd, accept_fd, epfd, err; + struct epoll_event ev; + socklen_t addrlen; + + addrlen = sizeof(struct sockaddr_in6); + ev.events = EPOLLIN; + + epfd = epoll_create(1); + CHECK(epfd == -1, "epoll_create()", "epfd:%d errno:%d", epfd, errno); + + /* Prepare listen_fd */ + listen_fd = socket(AF_INET6, SOCK_STREAM | SOCK_NONBLOCK, 0); + CHECK(listen_fd == -1, "socket()", "listen_fd:%d errno:%d", + listen_fd, errno); + + init_loopback6(&srv_sa6); + err = bind(listen_fd, (struct sockaddr *)&srv_sa6, sizeof(srv_sa6)); + CHECK(err, "bind(listen_fd)", "err:%d errno:%d", err, errno); + + err = getsockname(listen_fd, (struct sockaddr *)&srv_sa6, &addrlen); + CHECK(err, "getsockname(listen_fd)", "err:%d errno:%d", err, errno); + + err = listen(listen_fd, 1); + CHECK(err, "listen(listen_fd)", "err:%d errno:%d", err, errno); + + /* Prepare cli_fd */ + cli_fd = socket(AF_INET6, SOCK_STREAM | SOCK_NONBLOCK, 0); + CHECK(cli_fd == -1, "socket()", "cli_fd:%d errno:%d", cli_fd, errno); + + init_loopback6(&cli_sa6); + err = bind(cli_fd, (struct sockaddr *)&cli_sa6, sizeof(cli_sa6)); + CHECK(err, "bind(cli_fd)", "err:%d errno:%d", err, errno); + + err = getsockname(cli_fd, (struct sockaddr *)&cli_sa6, &addrlen); + CHECK(err, "getsockname(cli_fd)", "err:%d errno:%d", + err, errno); + + /* Update addr_map with srv_sa6 and cli_sa6 */ + err = bpf_map_update_elem(addr_map_fd, &srv_idx, &srv_sa6, 0); + CHECK(err, "map_update", "err:%d errno:%d", err, errno); + + err = bpf_map_update_elem(addr_map_fd, &cli_idx, &cli_sa6, 0); + CHECK(err, "map_update", "err:%d errno:%d", err, errno); + + /* Connect from cli_sa6 to srv_sa6 */ + err = connect(cli_fd, (struct sockaddr *)&srv_sa6, addrlen); + printf("srv_sa6.sin6_port:%u cli_sa6.sin6_port:%u\n\n", + ntohs(srv_sa6.sin6_port), ntohs(cli_sa6.sin6_port)); + CHECK(err && errno != EINPROGRESS, + "connect(cli_fd)", "err:%d errno:%d", err, errno); + + ev.data.fd = listen_fd; + err = epoll_ctl(epfd, EPOLL_CTL_ADD, listen_fd, &ev); + CHECK(err, "epoll_ctl(EPOLL_CTL_ADD, listen_fd)", "err:%d errno:%d", + err, errno); + + /* Accept the connection */ + /* Have some timeout in accept(listen_fd). Just in case. */ + err = epoll_wait(epfd, &ev, 1, 1000); + CHECK(err != 1 || ev.data.fd != listen_fd, + "epoll_wait(listen_fd)", + "err:%d errno:%d ev.data.fd:%d listen_fd:%d", + err, errno, ev.data.fd, listen_fd); + + accept_fd = accept(listen_fd, NULL, NULL); + CHECK(accept_fd == -1, "accept(listen_fd)", "accept_fd:%d errno:%d", + accept_fd, errno); + close(listen_fd); + + /* Send some data from accept_fd to cli_fd */ + err = send(accept_fd, DATA, DATA_LEN, 0); + CHECK(err != DATA_LEN, "send(accept_fd)", "err:%d errno:%d", + err, errno); + + /* Have some timeout in recv(cli_fd). Just in case. */ + ev.data.fd = cli_fd; + err = epoll_ctl(epfd, EPOLL_CTL_ADD, cli_fd, &ev); + CHECK(err, "epoll_ctl(EPOLL_CTL_ADD, cli_fd)", "err:%d errno:%d", + err, errno); + + err = epoll_wait(epfd, &ev, 1, 1000); + CHECK(err != 1 || ev.data.fd != cli_fd, + "epoll_wait(cli_fd)", "err:%d errno:%d ev.data.fd:%d cli_fd:%d", + err, errno, ev.data.fd, cli_fd); + + err = recv(cli_fd, NULL, 0, MSG_TRUNC); + CHECK(err, "recv(cli_fd)", "err:%d errno:%d", err, errno); + + close(epfd); + close(accept_fd); + close(cli_fd); + + check_result(); +} + +int main(int argc, char **argv) +{ + struct bpf_prog_load_attr attr = { + .file = "test_sock_fields_kern.o", + .prog_type = BPF_PROG_TYPE_CGROUP_SKB, + .expected_attach_type = BPF_CGROUP_INET_EGRESS, + }; + int cgroup_fd, prog_fd, err; + struct bpf_object *obj; + struct bpf_map *map; + + err = setup_cgroup_environment(); + CHECK(err, "setup_cgroup_environment()", "err:%d errno:%d", + err, errno); + + atexit(cleanup_cgroup_environment); + + /* Create a cgroup, get fd, and join it */ + cgroup_fd = create_and_get_cgroup(TEST_CGROUP); + CHECK(cgroup_fd == -1, "create_and_get_cgroup()", + "cgroup_fd:%d errno:%d", cgroup_fd, errno); + + err = join_cgroup(TEST_CGROUP); + CHECK(err, "join_cgroup", "err:%d errno:%d", err, errno); + + err = bpf_prog_load_xattr(&attr, &obj, &prog_fd); + CHECK(err, "bpf_prog_load_xattr()", "err:%d", err); + + err = bpf_prog_attach(prog_fd, cgroup_fd, BPF_CGROUP_INET_EGRESS, 0); + CHECK(err == -1, "bpf_prog_attach(CPF_CGROUP_INET_EGRESS)", + "err:%d errno%d", err, errno); + close(cgroup_fd); + + map = bpf_object__find_map_by_name(obj, "addr_map"); + CHECK(!map, "cannot find addr_map", "(null)"); + addr_map_fd = bpf_map__fd(map); + + map = bpf_object__find_map_by_name(obj, "sock_result_map"); + CHECK(!map, "cannot find sock_result_map", "(null)"); + sk_map_fd = bpf_map__fd(map); + + map = bpf_object__find_map_by_name(obj, "tcp_sock_result_map"); + CHECK(!map, "cannot find tcp_sock_result_map", "(null)"); + tp_map_fd = bpf_map__fd(map); + + map = bpf_object__find_map_by_name(obj, "linum_map"); + CHECK(!map, "cannot find linum_map", "(null)"); + linum_map_fd = bpf_map__fd(map); + + test(); + + bpf_object__close(obj); + cleanup_cgroup_environment(); + + printf("PASS\n"); + + return 0; +} diff --git a/tools/testing/selftests/bpf/test_sock_fields_kern.c b/tools/testing/selftests/bpf/test_sock_fields_kern.c new file mode 100644 index 000000000000..de1a43e8f610 --- /dev/null +++ b/tools/testing/selftests/bpf/test_sock_fields_kern.c @@ -0,0 +1,152 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Copyright (c) 2019 Facebook */ + +#include +#include +#include + +#include "bpf_helpers.h" +#include "bpf_endian.h" + +enum bpf_array_idx { + SRV_IDX, + CLI_IDX, + __NR_BPF_ARRAY_IDX, +}; + +struct bpf_map_def SEC("maps") addr_map = { + .type = BPF_MAP_TYPE_ARRAY, + .key_size = sizeof(__u32), + .value_size = sizeof(struct sockaddr_in6), + .max_entries = __NR_BPF_ARRAY_IDX, +}; + +struct bpf_map_def SEC("maps") sock_result_map = { + .type = BPF_MAP_TYPE_ARRAY, + .key_size = sizeof(__u32), + .value_size = sizeof(struct bpf_sock), + .max_entries = __NR_BPF_ARRAY_IDX, +}; + +struct bpf_map_def SEC("maps") tcp_sock_result_map = { + .type = BPF_MAP_TYPE_ARRAY, + .key_size = sizeof(__u32), + .value_size = sizeof(struct bpf_tcp_sock), + .max_entries = __NR_BPF_ARRAY_IDX, +}; + +struct bpf_map_def SEC("maps") linum_map = { + .type = BPF_MAP_TYPE_ARRAY, + .key_size = sizeof(__u32), + .value_size = sizeof(__u32), + .max_entries = 1, +}; + +static bool is_loopback6(__u32 *a6) +{ + return !a6[0] && !a6[1] && !a6[2] && a6[3] == bpf_htonl(1); +} + +static void skcpy(struct bpf_sock *dst, + const struct bpf_sock *src) +{ + dst->bound_dev_if = src->bound_dev_if; + dst->family = src->family; + dst->type = src->type; + dst->protocol = src->protocol; + dst->mark = src->mark; + dst->priority = src->priority; + dst->src_ip4 = src->src_ip4; + dst->src_ip6[0] = src->src_ip6[0]; + dst->src_ip6[1] = src->src_ip6[1]; + dst->src_ip6[2] = src->src_ip6[2]; + dst->src_ip6[3] = src->src_ip6[3]; + dst->src_port = src->src_port; + dst->dst_ip4 = src->dst_ip4; + dst->dst_ip6[0] = src->dst_ip6[0]; + dst->dst_ip6[1] = src->dst_ip6[1]; + dst->dst_ip6[2] = src->dst_ip6[2]; + dst->dst_ip6[3] = src->dst_ip6[3]; + dst->dst_port = src->dst_port; + dst->state = src->state; +} + +static void tpcpy(struct bpf_tcp_sock *dst, + const struct bpf_tcp_sock *src) +{ + dst->snd_cwnd = src->snd_cwnd; + dst->srtt_us = src->srtt_us; + dst->rtt_min = src->rtt_min; + dst->snd_ssthresh = src->snd_ssthresh; + dst->rcv_nxt = src->rcv_nxt; + dst->snd_nxt = src->snd_nxt; + dst->snd_una = src->snd_una; + dst->mss_cache = src->mss_cache; + dst->ecn_flags = src->ecn_flags; + dst->rate_delivered = src->rate_delivered; + dst->rate_interval_us = src->rate_interval_us; + dst->packets_out = src->packets_out; + dst->retrans_out = src->retrans_out; + dst->total_retrans = src->total_retrans; + dst->segs_in = src->segs_in; + dst->data_segs_in = src->data_segs_in; + dst->segs_out = src->segs_out; + dst->data_segs_out = src->data_segs_out; + dst->lost_out = src->lost_out; + dst->sacked_out = src->sacked_out; + dst->bytes_received = src->bytes_received; + dst->bytes_acked = src->bytes_acked; +} + +#define RETURN { \ + linum = __LINE__; \ + bpf_map_update_elem(&linum_map, &idx0, &linum, 0); \ + return 1; \ +} + +SEC("cgroup_skb/egress") +int read_sock_fields(struct __sk_buff *skb) +{ + __u32 srv_idx = SRV_IDX, cli_idx = CLI_IDX, idx; + struct sockaddr_in6 *srv_sa6, *cli_sa6; + struct bpf_tcp_sock *tp, *tp_ret; + struct bpf_sock *sk, *sk_ret; + __u32 linum, idx0 = 0; + + sk = skb->sk; + if (!sk || sk->state == 10) + RETURN; + + sk = bpf_sk_fullsock(sk); + if (!sk || sk->family != AF_INET6 || sk->protocol != IPPROTO_TCP || + !is_loopback6(sk->src_ip6)) + RETURN; + + tp = bpf_tcp_sock(sk); + if (!tp) + RETURN; + + srv_sa6 = bpf_map_lookup_elem(&addr_map, &srv_idx); + cli_sa6 = bpf_map_lookup_elem(&addr_map, &cli_idx); + if (!srv_sa6 || !cli_sa6) + RETURN; + + if (sk->src_port == bpf_ntohs(srv_sa6->sin6_port)) + idx = srv_idx; + else if (sk->src_port == bpf_ntohs(cli_sa6->sin6_port)) + idx = cli_idx; + else + RETURN; + + sk_ret = bpf_map_lookup_elem(&sock_result_map, &idx); + tp_ret = bpf_map_lookup_elem(&tcp_sock_result_map, &idx); + if (!sk_ret || !tp_ret) + RETURN; + + skcpy(sk_ret, sk); + tpcpy(tp_ret, tp); + + RETURN; +} + +char _license[] SEC("license") = "GPL"; -- cgit v1.2.3-59-g8ed1b