Re: [PATCH v2 bpf-next 4/6] selftests/bpf: Add IPv4 and IPv6 sockaddr test cases

On 4/12/24 9:52 AM, Jordan Rife wrote:

This patch lays the groundwork for testing IPv4 and IPv6 sockaddr hooks and their interaction with both socket syscalls and kernel functions (e.g. kernel_connect, kernel_bind, etc.) and moves the test cases from the old-style bpf/test_sock_addr.c self test into the sock_addr prog_test.

Thanks for moving the existing sock_addr test to the test_progs.

+static int ping_once(int ipv, const char *addr) +{

• const char *ping_cmd_prefix = "ping -";
• if (!SYS_NOFAIL("type ping%d >/dev/null 2>&1", ipv))

• ping_cmd_prefix = "ping";
  

• return SYS_NOFAIL("%s%d -q -c 1 -W 1 %s >/dev/null 2>&1",

• 	  ping_cmd_prefix, ipv, addr);
  

+}

+static int setup_test_env(void) +{

• SYS(err, "ip link add dev %s1 type veth peer name %s2", TEST_IF_PREFIX,

•    TEST_IF_PREFIX);
  

I would like to take this chance to simplify the setup.

Does it need a veth pair? The %s2 interface is not used.

Can it be done in lo alone?

Also, all this setup (and test) has to be done in a new netns. Anything blocking the kfunc in patch 2 using the current task netns instead of the init_net?

• SYS(err, "ip link set %s1 up", TEST_IF_PREFIX);
• SYS(err, "ip link set %s2 up", TEST_IF_PREFIX);
• SYS(err, "ip -4 addr add %s/8 dev %s1", TEST_IPV4, TEST_IF_PREFIX);
• SYS(err, "ip -6 addr add %s/128 dev %s1", TEST_IPV6, TEST_IF_PREFIX);

Add nodad to the "ip -6 addr add...". just in case it may add unnecessary delay.

• int i;
• for (i = 0; i < 5; i++) {

• if (!ping_once(4, TEST_IPV4) && !ping_once(6, TEST_IPV6))
  

This interface/address ping should not be needed. Other tests under prog_tests/ don't need this interface/address ping also.

• 	return 0;
  

• }
• ASSERT_FAIL("Timed out waiting for test IP to become available.");

+err:

• return -1;

+}

+static void cleanup_test_env(void) +{

• SYS_NOFAIL("ip link del %s1 2>/dev/null", TEST_IF_PREFIX);
• SYS_NOFAIL("ip link del %s2 2>/dev/null", TEST_IF_PREFIX);

Using lo will make this easier. Regardless, the link should go away with the netns.

+}

• void test_sock_addr(void) { int cgroup_fd = -1; void *skel;

• if (!ASSERT_OK(setup_test_env(), "setup_test_env"))

This will probably have to be called after the test__join_cgroup() if it also creates a new netns.

pw-bot: cr

• goto cleanup;
  

• cgroup_fd = test__join_cgroup("/sock_addr"); if (!ASSERT_GE(cgroup_fd, 0, "join_cgroup")) goto cleanup;

@@ -609,4 +755,5 @@ void test_sock_addr(void) cleanup: if (cgroup_fd >= 0) close(cgroup_fd);

• cleanup_test_env(); }