Re: [PATCH bpf-next 3/3] bpf: Update LSM selftests for bpf_ima_inode_hash

On Fri, Nov 20, 2020 at 7:11 PM Yonghong Song [off-list ref] wrote:

On 11/20/20 5:17 AM, KP Singh wrote:

quoted

From: KP Singh <redacted>

- Update the IMA policy before executing the test binary (this is not an
   override of the policy, just an append that ensures that hashes are
   calculated on executions).

- Call the bpf_ima_inode_hash in the bprm_committed_creds hook and check
   if the call succeeded and a hash was calculated.

Signed-off-by: KP Singh <redacted>

LGTM with a few nits below.

Acked-by: Yonghong Song <redacted>

quoted

---
  tools/testing/selftests/bpf/config            |  3 ++

[...]

quoted

  }

[...]

quoted

+
  void test_test_lsm(void)
  {
      struct lsm *skel = NULL;

@@ -66,6 +88,10 @@ void test_test_lsm(void)
      if (CHECK(err, "attach", "lsm attach failed: %d\n", err))
              goto close_prog;

+     err = update_ima_policy();
+     if (CHECK(err != 0, "update_ima_policy", "error = %d\n", err))
+             goto close_prog;

"err != 0" => err?
"error = %d" => "err %d" for consistency with other usage in this function.

Done.

quoted

+
      err = exec_cmd(&skel->bss->monitored_pid);
      if (CHECK(err < 0, "exec_cmd", "err %d errno %d\n", err, errno))
              goto close_prog;

@@ -83,6 +109,12 @@ void test_test_lsm(void)

[...]

quoted

  int mprotect_count = 0;
  int bprm_count = 0;
+int ima_hash_ret = -1;

The helper returns type "long", but "int" type here should be fine too.

Changed it to long for correctness.