lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Thu, 26 Oct 2023 15:06:39 +0800
From:   David Gow <davidgow@...gle.com>
To:     Maxime Ripard <mripard@...nel.org>
Cc:     Brendan Higgins <brendan.higgins@...ux.dev>,
        Jani Nikula <jani.nikula@...ux.intel.com>,
        Rae Moar <rmoar@...gle.com>, linux-kselftest@...r.kernel.org,
        kunit-dev@...glegroups.com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] kunit: Warn if tests are slow

On Wed, 20 Sept 2023 at 16:49, Maxime Ripard <mripard@...nel.org> wrote:
>
> Kunit recently gained support to setup attributes, the first one being
> the speed of a given test, then allowing to filter out slow tests.
>
> A slow test is defined in the documentation as taking more than one
> second. There's an another speed attribute called "super slow" but whose
> definition is less clear.
>
> Add support to the test runner to check the test execution time, and
> report tests that should be marked as slow but aren't.
>
> Signed-off-by: Maxime Ripard <mripard@...nel.org>
>
> ---
>
> To: Brendan Higgins <brendan.higgins@...ux.dev>
> To: David Gow <davidgow@...gle.com>
> Cc: Jani Nikula <jani.nikula@...ux.intel.com>
> Cc: Rae Moar <rmoar@...gle.com>
> Cc: linux-kselftest@...r.kernel.org
> Cc: kunit-dev@...glegroups.com
> Cc: linux-kernel@...r.kernel.org
>
> Changes from v1:
> - Split the patch out of the series
> - Change to trigger the warning only if the runtime is twice the
>   threshold (Jani, Rae)
> - Split the speed check into a separate function (Rae)
> - Link: https://lore.kernel.org/all/20230911-kms-slow-tests-v1-0-d3800a69a1a1@kernel.org/
> ---

I quite like this, though agree somewhat with Rae's comments below.

I personally think the time thresholds are, by necessity, very
'fuzzy', due to the varying speeds of different hardware. Fortunately,
the actual runtime of tests seems pretty well stratified, so the exact
threshold doesn't really matter much.

I ran some tests here, and all of the tests currently not marked slow
take <1s on every machine I tried (including the ancient 66MHz 486),
except for the drm_mm_* ones (which takes ~6s on my laptop, and times
out after 15 minutes on the aforementioned 486). Both the 1s and 2s
timeouts successfully distinguish those cases.

Ideally, I think we'd have something like:
#define KUNIT_SPEED_SLOW_THRESHOLD_S 1 /* 1 sec threshold for 'slow' tests */
#define KUNIT_SPEED_WARNING_MULTIPLIER 2 /* Warn when a test takes >
twice the threshold. */
#define KUNIT_SPEED_SLOW_WARNING_THRESHOLD_S
(KUNIT_SPEED_WARNING_MULTIPLIER * KUNIT_SPEED_SLOW_THRESHOLD_S)

Which is perhaps excessively verbose, but is very clear as to what
we're doing. It also gives more scope to allow the ratio to be
configured for people with very slow / fast machines in the future.

Thoughts?

Otherwise, this looks good to me.

Reviewed-by: David Gow <davidgow@...gle.com>

Cheers,
-- David

>  lib/kunit/test.c | 27 +++++++++++++++++++++++++++
>  1 file changed, 27 insertions(+)
>
> diff --git a/lib/kunit/test.c b/lib/kunit/test.c
> index 49698a168437..a1d5dd2bf87d 100644
> --- a/lib/kunit/test.c
> +++ b/lib/kunit/test.c
> @@ -372,6 +372,25 @@ void kunit_init_test(struct kunit *test, const char *name, char *log)
>  }
>  EXPORT_SYMBOL_GPL(kunit_init_test);
>
> +#define KUNIT_SPEED_SLOW_THRESHOLD_S   1
> +
> +static void kunit_run_case_check_speed(struct kunit *test,
> +                                      struct kunit_case *test_case,
> +                                      struct timespec64 duration)
> +{
> +       enum kunit_speed speed = test_case->attr.speed;
> +
> +       if (duration.tv_sec < (2 * KUNIT_SPEED_SLOW_THRESHOLD_S))
> +               return;
> +
> +       if (speed == KUNIT_SPEED_VERY_SLOW || speed == KUNIT_SPEED_SLOW)
> +               return;
> +
> +       kunit_warn(test,
> +                  "Test should be marked slow (runtime: %lld.%09lds)",
> +                  duration.tv_sec, duration.tv_nsec);
> +}
> +
>  /*
>   * Initializes and runs test case. Does not clean up or do post validations.
>   */
> @@ -379,6 +398,8 @@ static void kunit_run_case_internal(struct kunit *test,
>                                     struct kunit_suite *suite,
>                                     struct kunit_case *test_case)
>  {
> +       struct timespec64 start, end;
> +
>         if (suite->init) {
>                 int ret;
>
> @@ -390,7 +411,13 @@ static void kunit_run_case_internal(struct kunit *test,
>                 }
>         }
>
> +       ktime_get_ts64(&start);
> +
>         test_case->run_case(test);
> +
> +       ktime_get_ts64(&end);
> +
> +       kunit_run_case_check_speed(test, test_case, timespec64_sub(end, start));
>  }
>
>  static void kunit_case_internal_cleanup(struct kunit *test)
> --
> 2.41.0
>

Download attachment "smime.p7s" of type "application/pkcs7-signature" (4003 bytes)

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ