Received-SPF: pass (google.com: domain of linux-kernel+bounces-148788-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Date: Wed, 17 Apr 2024 08:22:22 -0700
From: Nathan Chancellor <nathan@kernel.org>
To: Christian Brauner <brauner@kernel.org>
Cc: linux-kselftest@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: tools/testing/selftests/clone3/clone3_set_tid.c appears to always
 pass?
Message-ID: <20240417152222.GA1517581@dev-arch.thelio-3990X>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

Hi Christian,

I am looking at tools/testing/selftests/clone3/clone3_set_tid.c as part
of a patch to clean up the uses of 'return ksft_exit_...();' throughout
the selftests (as they call exit() so they do not return) and I noticed
that it seems to always pass even when there may have been an error?

      if (waitpid(ns_pid, &status, 0) < 0) {
          ksft_print_msg("Child returned %s\n", strerror(errno));
          ret = -errno;
          goto out;
      }

      ...
  out:
      ret = 0;

      return !ret ? ksft_exit_pass() : ksft_exit_fail();
  }

Should the ret and out label positions be switched? Alternatively, it
seems like ret and out do not have that many uses, perhaps it would just
be better to call the ksft_exit_...() directly in their respective
paths? I am not going to touch it as part of my patch but I felt it was
worth reporting since it appears to have been there since the
introduction of this test in commit 41585bbeeef9 ("selftests: add tests
for clone3() with *set_tid").

Cheers,
Nathan