From: Amery Hung <amery.hung@xxxxxxxxxxxxx>
If eBPF users mistakenly reserve smaller header options than the actual
size in BPF_SOCK_OPS_HDR_OPT_LEN_CB, bpf_reserve_hdr_opt should return an
appropriate error value, and there will be no packet dropping.
Signed-off-by: Zijian Zhang <zijianzhang@xxxxxxxxxxxxx>
Signed-off-by: Amery Hung <amery.hung@xxxxxxxxxxxxx>
---
.../bpf/prog_tests/tcp_hdr_options.c | 51 +++++++++++++
.../bpf/progs/test_reserve_tcp_hdr_options.c | 71 +++++++++++++++++++
2 files changed, 122 insertions(+)
create mode 100644 tools/testing/selftests/bpf/progs/test_reserve_tcp_hdr_options.c
diff --git a/tools/testing/selftests/bpf/prog_tests/tcp_hdr_options.c b/tools/testing/selftests/bpf/prog_tests/tcp_hdr_options.c
index 56685fc03c7e..9c250b5bf00a 100644
--- a/tools/testing/selftests/bpf/prog_tests/tcp_hdr_options.c
+++ b/tools/testing/selftests/bpf/prog_tests/tcp_hdr_options.c
@@ -14,6 +14,7 @@
#include "test_tcp_hdr_options.h"
#include "test_tcp_hdr_options.skel.h"
#include "test_misc_tcp_hdr_options.skel.h"
+#include "test_reserve_tcp_hdr_options.skel.h"
#define LO_ADDR6 "::1"
#define CG_NAME "/tcpbpf-hdr-opt-test"
@@ -25,6 +26,7 @@ static struct bpf_test_option exp_active_fin_in;
static struct hdr_stg exp_passive_hdr_stg;
static struct hdr_stg exp_active_hdr_stg = { .active = true, };
+static struct test_reserve_tcp_hdr_options *reserve_skel;
static struct test_misc_tcp_hdr_options *misc_skel;
static struct test_tcp_hdr_options *skel;
static int lport_linum_map_fd;
@@ -513,6 +515,49 @@ static void misc(void)
bpf_link__destroy(link);
}
+static void reserve_hdr_opt(void)
+{
+ struct bpf_link *link;
+ struct sk_fds sk_fds;
+ char send_msg[1500];
+ char recv_msg[sizeof(send_msg)];
+ int ret;
+
+ if (!ASSERT_OK(system("ip link set dev lo mtu 1500"), "set dev lo mtu to 1500"))
+ return;
+
+ lport_linum_map_fd = bpf_map__fd(reserve_skel->maps.lport_linum_map);
+
+ link = bpf_program__attach_cgroup(reserve_skel->progs.reserve_tcp_hdr_options, cg_fd);
+ if (!ASSERT_OK_PTR(link, "attach_cgroup(reserve_tcp_hdr_options)"))
+ return;
+
+ if (sk_fds_connect(&sk_fds, false)) {
+ bpf_link__destroy(link);
+ return;
+ }
+
+ ret = send(sk_fds.active_fd, send_msg, sizeof(send_msg),
+ MSG_EOR);
+ if (!ASSERT_EQ(ret, sizeof(send_msg), "send(msg)"))
+ goto check_linum;
+
+ ret = read(sk_fds.passive_fd, recv_msg, sizeof(recv_msg));
+ if (!ASSERT_EQ(ret, sizeof(send_msg), "read(msg)"))
+ goto check_linum;
+
+ if (sk_fds_shutdown(&sk_fds))
+ goto check_linum;
+
+ ASSERT_FALSE(reserve_skel->bss->nr_err_reserve, "unexpected nr_err_reserve");
+ ASSERT_TRUE(reserve_skel->bss->nr_nospc, "unexpected nr_nospc");
+
+check_linum:
+ ASSERT_FALSE(check_error_linum(&sk_fds), "check_error_linum");
+ sk_fds_close(&sk_fds);
+ bpf_link__destroy(link);
+}
+
struct test {
const char *desc;
void (*run)(void);
@@ -526,6 +571,7 @@ static struct test tests[] = {
DEF_TEST(fastopen_estab),
DEF_TEST(fin),
DEF_TEST(misc),
+ DEF_TEST(reserve_hdr_opt),
};
void test_tcp_hdr_options(void)
@@ -540,6 +586,10 @@ void test_tcp_hdr_options(void)
if (!ASSERT_OK_PTR(misc_skel, "open and load misc test skel"))
goto skel_destroy;
+ reserve_skel = test_reserve_tcp_hdr_options__open_and_load();
+ if (!ASSERT_OK_PTR(reserve_skel, "open and load reserve test skel"))
+ goto skel_destroy;
+
cg_fd = test__join_cgroup(CG_NAME);
if (!ASSERT_GE(cg_fd, 0, "join_cgroup"))
goto skel_destroy;
@@ -558,6 +608,7 @@ void test_tcp_hdr_options(void)
close(cg_fd);
skel_destroy:
+ test_reserve_tcp_hdr_options__destroy(reserve_skel);
test_misc_tcp_hdr_options__destroy(misc_skel);
test_tcp_hdr_options__destroy(skel);
}
diff --git a/tools/testing/selftests/bpf/progs/test_reserve_tcp_hdr_options.c b/tools/testing/selftests/bpf/progs/test_reserve_tcp_hdr_options.c
new file mode 100644
index 000000000000..a40d31c4ae1b
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/test_reserve_tcp_hdr_options.c
@@ -0,0 +1,71 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2024 ByteDance Inc. */
+
+#include <stddef.h>
+#include <errno.h>
+#include <stdbool.h>
+#include <sys/types.h>
+#include <sys/socket.h>
+#include <linux/ipv6.h>
+#include <linux/tcp.h>
+#include <linux/socket.h>
+#include <linux/bpf.h>
+#include <linux/types.h>
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_endian.h>
+#define BPF_PROG_TEST_TCP_HDR_OPTIONS
+#include "test_tcp_hdr_options.h"
+
+unsigned int nr_err_reserve = 0;
+unsigned int nr_nospc = 0;
+
+static bool skops_current_mss(const struct bpf_sock_ops *skops)
+{
+ return skops->args[0] == BPF_WRITE_HDR_TCP_CURRENT_MSS;
+}
+
+static int handle_hdr_opt_len(struct bpf_sock_ops *skops)
+{
+ int err;
+
+ if (skops_current_mss(skops)) {
+ err = bpf_reserve_hdr_opt(skops, 4, 0);
+ if (err) {
+ nr_err_reserve++;
+ RET_CG_ERR(err);
+ }
+ } else {
+ err = bpf_reserve_hdr_opt(skops, 8, 0);
+ if (err) {
+ if (err == -ENOSPC) {
+ nr_nospc++;
+ } else {
+ nr_err_reserve++;
+ RET_CG_ERR(err);
+ }
+ }
+ }
+
+ return CG_OK;
+}
+
+SEC("sockops")
+int reserve_tcp_hdr_options(struct bpf_sock_ops *skops)
+{
+ switch (skops->op) {
+ case BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB:
+ case BPF_SOCK_OPS_ACTIVE_ESTABLISHED_CB:
+ bpf_sock_ops_cb_flags_set(skops,
+ skops->bpf_sock_ops_cb_flags |
+ BPF_SOCK_OPS_WRITE_HDR_OPT_CB_FLAG);
+ break;
+ case BPF_SOCK_OPS_HDR_OPT_LEN_CB:
+ return handle_hdr_opt_len(skops);
+ case BPF_SOCK_OPS_WRITE_HDR_OPT_CB:
+ break;
+ }
+
+ return CG_OK;
+}
+
+char _license[] SEC("license") = "GPL";
--
2.20.1
