> diff --git a/tools/testing/selftests/bpf/network_helpers.c
> b/tools/testing/selftests/bpf/network_helpers.c
> index b82f572641b7..2f2719a28f02 100644
> --- a/tools/testing/selftests/bpf/network_helpers.c
> +++ b/tools/testing/selftests/bpf/network_helpers.c
[ ... ]
> @@ -40,6 +41,8 @@
> #define IPPROTO_MPTCP 262
> #endif
>
> +#define CONNECTION_IN_PROGRESS_TIMEOUT_MS 3000
> +
> #define clean_errno() (errno == 0 ? "None" : strerror(errno))
> #define log_err(MSG, ...) ({ \
> int __save = errno; \
[ ... ]
> @@ -294,7 +297,8 @@ int client_socket(int family, int type,
> int connect_to_addr(int type, const struct sockaddr_storage *addr, socklen_t
> addrlen,
> const struct network_helper_opts *opts)
> {
> - int fd;
> + socklen_t errlen;
> + int fd, err;
>
> if (!opts)
> opts = &default_opts;
> @@ -305,13 +309,42 @@ int connect_to_addr(int type, const struct
> sockaddr_storage *addr, socklen_t add
> return -1;
> }
>
> - if (connect(fd, (const struct sockaddr *)addr, addrlen)) {
> + err = connect(fd, (const struct sockaddr *)addr, addrlen);
> + if (err && errno == EINPROGRESS) {
> + struct pollfd pfd = { .fd = fd, .events = POLLOUT };
> +
> + err = poll(&pfd, 1, CONNECTION_IN_PROGRESS_TIMEOUT_MS);
Does the fixed 3000 ms timeout override opts->timeout_ms correctly?
client_socket() in this file passes opts->timeout_ms to settimeo(), which
converts it into SO_RCVTIMEO and SO_SNDTIMEO:
tools/testing/selftests/bpf/network_helpers.c:settimeo() {
...
if (timeout_ms > 0) {
timeout.tv_sec = timeout_ms / 1000;
timeout.tv_usec = (timeout_ms % 1000) * 1000;
} else {
timeout.tv_sec = 3;
timeout.tv_usec = 0;
}
...
}
tools/testing/selftests/bpf/network_helpers.c:client_socket() {
...
if (settimeo(fd, opts->timeout_ms))
goto error_close;
...
}
That opts->timeout_ms is the caller's requested timeout contract. But the
new poll() discards it and uses CONNECTION_IN_PROGRESS_TIMEOUT_MS instead,
which is always 3000.
This changes observable timing for callers in both directions:
Callers that request a SHORT timeout now wait longer than they asked for.
prog_tests/assign_reuse.c calls connect_to_fd(fd_sv, 100) at lines 42 and
66, requesting a 100 ms budget. A stalled connect now blocks for 100 ms
(SO_SNDTIMEO) plus 3000 ms in poll() - a 30x overrun.
prog_tests/test_tc_tunnel.c sets .timeout_ms = 1000 at line 171.
The commit message explicitly argues against lengthening timeouts:
"increasing the timeout value would increase overall test execution
duration again", yet the patch lengthens these callers' timeouts from
their requested values.
Callers that request a LONG timeout get a shorter window than they asked
for. prog_tests/tc_redirect.c uses TIMEOUT_MILLIS (10000) at lines 428,
561, and 658. prog_tests/xdp_synproxy.c uses 10000 at line 135. On a
loaded runner the connect is now abandoned at SO_SNDTIMEO + 3 s rather
than the requested 10 s.
Would deriving the poll timeout from opts->timeout_ms preserve the single
timeout contract that callers already use?
>
> + if (err <= 0) {
> + if (err == 0) {
> + log_err("Connection timeout");
> + errno = ETIMEDOUT;
> + } else {
> + log_err("Failed to poll connection status");
> + }
> + goto close;
> + }
> +
> + errlen = sizeof(err);
> + if (getsockopt(fd, SOL_SOCKET, SO_ERROR, &err, &errlen) < 0) {
> + log_err("Failed to getsockopt");
> + goto close;
> + }
> +
> + if (err) {
> + log_err("Eventually failed to connect to server");
> + errno = err;
> + goto close;
> + }
> + } else if (err) {
> log_err("Failed to connect to server");
> - save_errno_close(fd);
> - return -1;
> + goto close;
> }
>
> return fd;
> +close:
> + save_errno_close(fd);
> + return -1;
> }
---
AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md
CI run summary: https://github.com/kernel-patches/bpf/actions/runs/30795380743
