On 5/25/24 5:08 AM, Geliang Tang wrote:
From: Geliang Tang tanggeliang@kylinos.cn
This patch uses connect_to_fd_opts() instead of using connect_fd_to_fd() and settcpca() in do_test() in prog_tests/bpf_tcp_ca.c to accept a struct network_helper_opts argument.
Then define a dctcp dedicated post_socket_cb callback stg_post_socket_cb(), invoking both cc_cb() and bpf_map_update_elem() in it, and set it in test_dctcp(). For passing map_fd into stg_post_socket_cb() callback, a new member map_fd is added in struct cb_opts.
Signed-off-by: Geliang Tang tanggeliang@kylinos.cn
.../selftests/bpf/prog_tests/bpf_tcp_ca.c | 42 ++++++++++--------- 1 file changed, 22 insertions(+), 20 deletions(-)
diff --git a/tools/testing/selftests/bpf/prog_tests/bpf_tcp_ca.c b/tools/testing/selftests/bpf/prog_tests/bpf_tcp_ca.c index ebc7d4616880..9a7c3dc39008 100644 --- a/tools/testing/selftests/bpf/prog_tests/bpf_tcp_ca.c +++ b/tools/testing/selftests/bpf/prog_tests/bpf_tcp_ca.c @@ -25,6 +25,7 @@ static int expected_stg = 0xeB9F; struct cb_opts { const char *cc;
- int map_fd; };
static int settcpca(int fd, const char *tcp_ca) @@ -41,7 +42,6 @@ static int settcpca(int fd, const char *tcp_ca) static void do_test(const struct network_helper_opts *opts, const struct bpf_map *sk_stg_map) {
- struct cb_opts *cb_opts = (struct cb_opts *)opts->cb_opts; int lfd = -1, fd = -1; int err;
@@ -49,25 +49,9 @@ static void do_test(const struct network_helper_opts *opts, if (!ASSERT_NEQ(lfd, -1, "socket")) return;
- fd = socket(AF_INET6, SOCK_STREAM, 0);
- if (!ASSERT_NEQ(fd, -1, "socket")) {
close(lfd);return;- }
- if (settcpca(fd, cb_opts->cc))
goto done;- if (sk_stg_map) {
err = bpf_map_update_elem(bpf_map__fd(sk_stg_map), &fd,&expected_stg, BPF_NOEXIST);if (!ASSERT_OK(err, "bpf_map_update_elem(sk_stg_map)"))goto done;- }
- /* connect to server */
- err = connect_fd_to_fd(fd, lfd, 0);
- if (!ASSERT_NEQ(err, -1, "connect"))
- fd = connect_to_fd_opts(lfd, opts);
- if (!ASSERT_NEQ(fd, -1, "connect_to_fd_opts")) goto done;
if (sk_stg_map) { @@ -124,13 +108,30 @@ static void test_cubic(void) bpf_cubic__destroy(cubic_skel); } +static int stg_post_socket_cb(int fd, void *opts) +{
- struct cb_opts *cb_opts = (struct cb_opts *)opts;
- int err;
- err = settcpca(fd, cb_opts->cc);
- if (err)
return err;- err = bpf_map_update_elem(cb_opts->map_fd, &fd,
&expected_stg, BPF_NOEXIST);
Looking at it again, it is not very correct. At least this map update is unnecessary for the start_server_str() in do_test(). do_test() uses the same opts for start_server_str() and connect_to_fd_opts().
Also, the post_connect_cb in patch 8 is unnecessary for network_helpers.c. The connect_to_fd_opts helper is going to return the fd anyway. It is unnecessary to have the helper to do the post_connect_cb(fd). The caller (do_test here) can directly do that instead. Lets not add unnecessary post_connect_cb to the network_helpers.c now until there is a clear case that the caller cannot do it.
I would just add another "const struct network_helper_opts *cli_opts" to the do_test() to do what patch 7 and patch 8 need to do. Separate it from the server "opts".
Patach 1 to 5 is applied. Thanks.
- if (!ASSERT_OK(err, "bpf_map_update_elem(sk_stg_map)"))
return err;- return 0;
+}
- static void test_dctcp(void) { struct cb_opts cb_opts = { .cc = "bpf_dctcp", }; struct network_helper_opts opts = {
.post_socket_cb = cc_cb,
.post_socket_cb = stg_post_socket_cb,@@ -146,6 +147,7 @@ static void test_dctcp(void) return; }
- cb_opts.map_fd = bpf_map__fd(dctcp_skel->maps.sk_stg_map); do_test(&opts, dctcp_skel->maps.sk_stg_map); ASSERT_EQ(dctcp_skel->bss->stg_result, expected_stg, "stg_result");