On 10/17/24 3:49 PM, Jordan Rife wrote:
Move all BPF_CGROUP_INET6_POST_BIND and BPF_CGROUP_INET4_POST_BIND test
cases to a new prog_test, prog_tests/sock_post_bind.c, except for
LOAD_REJECT test cases.
Signed-off-by: Jordan Rife <jr...@google.com>
---
.../selftests/bpf/prog_tests/sock_post_bind.c | 417 ++++++++++++++++++
tools/testing/selftests/bpf/test_sock.c | 245 ----------
2 files changed, 417 insertions(+), 245 deletions(-)
create mode 100644 tools/testing/selftests/bpf/prog_tests/sock_post_bind.c
diff --git a/tools/testing/selftests/bpf/prog_tests/sock_post_bind.c
b/tools/testing/selftests/bpf/prog_tests/sock_post_bind.c
new file mode 100644
index 000000000000..c46537e3b9d4
--- /dev/null
+++ b/tools/testing/selftests/bpf/prog_tests/sock_post_bind.c
@@ -0,0 +1,417 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/bpf.h>
+#include <test_progs.h>
+#include "cgroup_helpers.h"
+
+static char bpf_log_buf[4096];
+static bool verbose;
How is verbose used and is it still needed?
[ ... ]
+ if (bind(sockfd, (const struct sockaddr *)&addr, len) == -1) {
+ /* sys_bind() may fail for different reasons, errno has to be
+ * checked to confirm that BPF program rejected it.
+ */
+ if (errno != EPERM)
+ goto err;
+ if (port_retry)
+ goto retry;
+ res = BIND_REJECT;
+ goto out;
+ }
[ ... ]
+void test_sock_post_bind(void)
+{
+ int cgroup_fd, i;
+
+ cgroup_fd = test__join_cgroup("/post_bind");
+ if (!ASSERT_GE(cgroup_fd, 0, "join_cgroup"))
nit. ASSERT_OK_FD().
Since the test binds to a specific ip/port, please run it in its own netns. It
is easy to do with netns_new and netns_free, a recent example:
https://lore.kernel.org/bpf/20241020-syncookie-v2-1-2db240225...@bootlin.com/
The same netns can be reused for different subtests of this "sock_post_bind"
test.
Others look good. Thanks for moving the test to test_progs.
pw-bot: cr
