Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230330220506.1399796-1-rmoar@google.com>
In-Reply-To: <20230330220506.1399796-1-rmoar@google.com>
From:   Daniel Latypov <dlatypov@google.com>
Date:   Thu, 30 Mar 2023 15:20:52 -0700
Message-ID: <CAGS_qxqNwVcymkG6-8Kv72oZc9aDqjFjBBmjr+f+mOVKT1bGvA@mail.gmail.com>
Subject: Re: [PATCH v1] kunit: add tests for using current KUnit test field
To:     Rae Moar <rmoar@google.com>
Cc:     brendanhiggins@google.com, davidgow@google.com,
        skhan@linuxfoundation.org, kunit-dev@googlegroups.com,
        linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

I've got a few minor comments below, but this otherwise looks good.
I like the idea of testing knuit_fail_current_test().


On Thu, Mar 30, 2023 at 3:05=E2=80=AFPM Rae Moar <rmoar@google.com> wrote:
>
> +static void kunit_current_kunit_test_field(struct kunit *test)
> +{
> +       struct kunit *current_test;
> +
> +       /* Check to ensure the result of current->kunit_test
> +        * is equivalent to current test.
> +        */
> +       current_test =3D current->kunit_test;
> +       KUNIT_EXPECT_PTR_EQ(test, test, current_test);

Perhaps we can combine this and the next test case down to
static void kunit_current_test(struct kunit *test) {
  /* There are two different ways of getting the current test */
  KUNIT_EXPECT_PTR_EQ(test, test, current->kunit_test);
  KUNIT_EXPECT_PTR_EQ(test, test, kunit_get_current_test());
}
?

> +}
> +
> +static void kunit_current_get_current_test(struct kunit *test)
> +{
> +       struct kunit *current_test1, *current_test2;
> +
> +       /* Check to ensure the result of kunit_get_current_test()
> +        * is equivalent to current test.
> +        */
> +       current_test1 =3D kunit_get_current_test();
> +       KUNIT_EXPECT_PTR_EQ(test, test, current_test1);
> +
> +       /* Check to ensure the result of kunit_get_current_test()
> +        * is equivalent to current->kunit_test.
> +        */
> +       current_test2 =3D current->kunit_test;
> +       KUNIT_EXPECT_PTR_EQ(test, current_test1, current_test2);

> +}
> +
> +static void kunit_current_fail_current_test(struct kunit *test)
> +{
> +       struct kunit fake;
> +
> +       /* Initialize fake test and set as current->kunit_test. */

Nit: I think the code is self-explanatory enough that we can drop this comm=
ent.

> +       kunit_init_test(&fake, "fake test", NULL);
> +       KUNIT_EXPECT_EQ(test, fake.status, KUNIT_SUCCESS);
> +       current->kunit_test =3D &fake;
> +
> +       /* Fail current test and expect status of fake test to be failed.=
 */

Nit: I think this comment could also be dropped or maybe shortened to
  kunit_fail_current_test("This should make `fake` fail");

or
  /* Now kunit_fail_current_test() should modify `fake`, not `test` */
  kunit_fail_current_test("This should make `fake` fail");

> +       kunit_fail_current_test("This test is supposed to fail.");
> +       KUNIT_EXPECT_EQ(test, fake.status, (enum kunit_status)KUNIT_FAILU=
RE);
> +

Hmm, should we try calling
  kunit_cleanup(&fake);
?

Right now this does resource cleanups, but we might have other state
to cleanup for our `fake` test object in the future.

Daniel