Re: [PATCH bpf-next v2 1/4] selftests/bpf: Use post_socket_cb in connect_to_fd_opts

On 5/14/24 9:20 PM, Geliang Tang wrote:

From: Geliang Tang tanggeliang@kylinos.cn

Since the post_socket_cb() callback is added in struct network_helper_opts, it's make sense to use it not only in __start_server(), but also in connect_to_fd_opts(). Then it can be used to set TCP_CONGESTION sockopt.

Add a post_socket_opts type member cb_opts into struct network_helper_opts, then cc can be moved into struct post_socket_opts from network_helper_opts. Define a new callback cc_cb() to set TCP_CONGESTION sockopt, and set it to post_socket_cb pointer.

Signed-off-by: Geliang Tang tanggeliang@kylinos.cn

tools/testing/selftests/bpf/network_helpers.c | 5 ++--- tools/testing/selftests/bpf/network_helpers.h | 6 ++++-- tools/testing/selftests/bpf/prog_tests/bpf_tcp_ca.c | 9 ++++++++- 3 files changed, 14 insertions(+), 6 deletions(-)

diff --git a/tools/testing/selftests/bpf/network_helpers.c b/tools/testing/selftests/bpf/network_helpers.c index 35250e6cde7f..d97f8a669b38 100644 --- a/tools/testing/selftests/bpf/network_helpers.c +++ b/tools/testing/selftests/bpf/network_helpers.c @@ -338,9 +338,8 @@ int connect_to_fd_opts(int server_fd, const struct network_helper_opts *opts) if (settimeo(fd, opts->timeout_ms)) goto error_close;

• if (opts->cc && opts->cc[0] &&

•    setsockopt(fd, SOL_TCP, TCP_CONGESTION, opts->cc,
  

•        strlen(opts->cc) + 1))
  

• if (opts->post_socket_cb &&

•    opts->post_socket_cb(fd, &opts->cb_opts))
  

  goto error_close;

if (!opts->noconnect) diff --git a/tools/testing/selftests/bpf/network_helpers.h b/tools/testing/selftests/bpf/network_helpers.h index 883c7ea9d8d5..e44a6e5d8344 100644 --- a/tools/testing/selftests/bpf/network_helpers.h +++ b/tools/testing/selftests/bpf/network_helpers.h @@ -21,16 +21,18 @@ typedef __u16 __sum16; #define VIP_NUM 5 #define MAGIC_BYTES 123 -struct post_socket_opts {}; +struct post_socket_opts {

• const char *cc;

+};

From looking its usage in this set, it has cc name and map fd in it. It becomes clear that it is not possible to have one generic/common "struct post_socket_opts" for all tests.

struct network_helper_opts {

• const char *cc; int timeout_ms; bool must_fail; bool noconnect; int type; int proto; int (*post_socket_cb)(int fd, const struct post_socket_opts *opts);

• struct post_socket_opts cb_opts;

It is better to have the individual test define its own callback opts struct. Something like this:

/* network_helpers.h */ struct network_helper_opts { /* ... */ int (*post_socket_cb)(int fd, void *opts); void *post_socket_opts; };

/* bpf_tcp_ca.c */ struct cb_opts { const char *cc; int map_fd; };

struct cb_opts cb_opts = { .cc = "bpf_dctcp", .map_fd = fd, };

struct network_helper_opts opts = { .post_socket_cb = stg_post_socket_cb, .post_sock_opts = &cb_opts, };

pw-bot: cr

}; /* ipv4 test vector */ 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 0aca02532794..9bc909fa0833 100644 --- a/tools/testing/selftests/bpf/prog_tests/bpf_tcp_ca.c +++ b/tools/testing/selftests/bpf/prog_tests/bpf_tcp_ca.c @@ -81,6 +81,12 @@ static void do_test(const char *tcp_ca, const struct bpf_map *sk_stg_map) close(fd); } +static int cc_cb(int fd, const struct post_socket_opts *opts) +{

• return setsockopt(fd, SOL_TCP, TCP_CONGESTION, opts->cc,

• 	  strlen(opts->cc) + 1);
  

+}

• static void test_cubic(void) { struct bpf_cubic *cubic_skel;

@@ -172,7 +178,8 @@ static void test_dctcp_fallback(void) { int err, lfd = -1, cli_fd = -1, srv_fd = -1; struct network_helper_opts opts = {

• .cc = "cubic",
  

• .cb_opts.cc = "cubic",
  

• .post_socket_cb = cc_cb,
  

  }; struct bpf_dctcp *dctcp_skel; struct bpf_link *link = NULL;