selftests/net: Add mixed select()+polling mode to TCP-AO tests

Currently, tcp_ao tests have two timeouts: TEST_RETRANSMIT_SEC and
TEST_TIMEOUT_SEC [by default 1 and 5 seconds]. The first one,
TEST_RETRANSMIT_SEC is used for operations that are expected to succeed
in order for a test to pass. It is usually not consumed and exists only
to avoid indefinite test run if the operation didn't complete.
The second one, TEST_RETRANSMIT_SEC exists for the tests that checking
operations, that are expected to fail/timeout. It is shorter as it is
fully consumed, with an expectation that if operation didn't succeed
during that period, it will timeout. And the related test that expects
the timeout is passing. The actual operation failure is then
cross-verified by other means like counters checks.

The issue with TEST_RETRANSMIT_SEC timeout is that 1 second is the exact
initial TCP timeout. So, in case the initial segment gets lost (quite
unlikely on local veth interface between two net namespaces, yet happens
in slow VMs), the retransmission never happens and as a result, the test
is not actually testing the functionality. Which in the end fails
counters checks.

As I want tcp_ao selftests to be fast and finishing in a reasonable
amount of time on manual run, I didn't consider increasing
TEST_RETRANSMIT_SEC.

Rather, initially, BPF_SOCK_OPS_TIMEOUT_INIT looked promising as a lever
to make the initial TCP timeout shorter. But as it's not a socket bpf
attached thing, but sock_ops (attaches to cgroups), the selftests would
have to use libbpf, which I wanted to avoid if not absolutely required.

Instead, use a mixed select() and counters polling mode with the longer
TEST_TIMEOUT_SEC timeout to detect running-away failed tests. It
actually not only allows losing segments and succeeding after
the previous TEST_RETRANSMIT_SEC timeout was consumed, but makes
the tests expecting timeout/failure pass faster.

The only test case taking longer (TEST_TIMEOUT_SEC) now is connect-deny
"wrong snd id", which checks for no key on SYN-ACK for which there is no
counter in the kernel (see tcp_make_synack()). Yet it can be speed up
by poking skpair from the trace event (see trace_tcp_ao_synack_no_key).

Fixes: ed9d09b309b1 ("selftests/net: Add a test for TCP-AO keys matching")
Reported-by: Jakub Kicinski <kuba@kernel.org>
Closes: https://lore.kernel.org/netdev/20241205070656.6ef344d7@kernel.org/
Signed-off-by: Dmitry Safonov <0x7f454c46@gmail.com>
Link: https://patch.msgid.link/20250319-tcp-ao-selftests-polling-v2-4-da48040153d1@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

1 files changed, 35 insertions, 2 deletions

diff --git a/tools/testing/selftests/net/tcp_ao/lib/aolib.h b/tools/testing/selftests/net/tcp_ao/lib/aolib.h
index b88d15df7688..3c8c5ac2a878 100644
--- a/tools/testing/selftests/net/tcp_ao/lib/aolib.h
+++ b/tools/testing/selftests/net/tcp_ao/lib/aolib.h
@@ -331,6 +331,7 @@ static inline int test_listen_socket(const union tcp_addr taddr,
  * If set to 0 - kernel will try to retransmit SYN number of times, set in
  * /proc/sys/net/ipv4/tcp_syn_retries
  * By default set to 1 to make tests pass faster on non-busy machine.
+ * [in process of removal, don't use in new tests]
  */
 #ifndef TEST_RETRANSMIT_SEC
 #define TEST_RETRANSMIT_SEC	1
@@ -483,8 +484,6 @@ static inline int test_set_ao_flags(int sk, bool ao_required, bool accept_icmps)
 }
 
 extern ssize_t test_server_run(int sk, ssize_t quota, time_t timeout_sec);
-extern ssize_t test_client_loop(int sk, char *buf, size_t buf_sz,
-				const size_t msg_len, time_t timeout_sec);
 extern int test_client_verify(int sk, const size_t msg_len, const size_t nr,
 			      time_t timeout_sec);
 
@@ -581,6 +580,40 @@ extern int test_assert_counters_key(const char *tst_name,
 extern void test_tcp_counters_free(struct tcp_counters *cnts);
 
 /*
+ * Polling for netns and socket counters during select()/connect() and also
+ * client/server messaging. Instead of constant timeout on underlying select(),
+ * check the counters and return early. This allows to pass the tests where
+ * timeout is expected without waiting for that fixing timeout (tests speed-up).
+ * Previously shorter timeouts were used for tests expecting to time out,
+ * but that leaded to sporadic false positives on counter checks failures,
+ * as one second timeouts aren't enough for TCP retransmit.
+ *
+ * Two sides of the socketpair (client/server) should synchronize failures
+ * using a shared variable *err, so that they can detect the other side's
+ * failure.
+ */
+extern int test_skpair_wait_poll(int sk, bool write, test_cnt cond,
+				 volatile int *err);
+extern int _test_skpair_connect_poll(int sk, const char *device,
+				     void *addr, size_t addr_sz,
+				     test_cnt cond, volatile int *err);
+static inline int test_skpair_connect_poll(int sk, const union tcp_addr taddr,
+					   unsigned int port,
+					   test_cnt cond, volatile int *err)
+{
+	sockaddr_af addr;
+
+	tcp_addr_to_sockaddr_in(&addr, &taddr, htons(port));
+	return _test_skpair_connect_poll(sk, veth_name,
+					 (void *)&addr, sizeof(addr), cond, err);
+}
+
+extern int test_skpair_client(int sk, const size_t msg_len, const size_t nr,
+			      test_cnt cond, volatile int *err);
+extern int test_skpair_server(int sk, ssize_t quota,
+			      test_cnt cond, volatile int *err);
+
+/*
  * Frees buffers allocated in test_get_tcp_counters().
  * The function doesn't expect new keys or keys removed between calls
  * to test_get_tcp_counters(). Check key counters manually if they