Re: [RFC PATCH v1 9/9] selftest: sync: disable tests that rely on not yet defined behaviour

[Shuah Khan]

On 03/09/2016 08:29 AM, Emilio LÃpez wrote:
> One of the tests rely on a behaviour only observed on the driver currently
> in use in Android. Disable it here until the behaviour is implemented
> or it is decided it should not be implemented on the driver in mainline.
> 
> Signed-off-by: Emilio LÃpez <emilio.lopez@xxxxxxxxxxxxxxx>
> 
> ---
> 
>  tools/testing/selftests/sync/sync_test.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/tools/testing/selftests/sync/sync_test.c b/tools/testing/selftests/sync/sync_test.c
> index c3e4c01..3f484318 100644
> --- a/tools/testing/selftests/sync/sync_test.c
> +++ b/tools/testing/selftests/sync/sync_test.c
> @@ -66,7 +66,14 @@ int main(void)
>  	err += RUN_TEST(test_fence_one_timeline_merge);
>  	err += RUN_TEST(test_fence_merge_same_fence);
>  	err += RUN_TEST(test_fence_multi_timeline_wait);
> +#if 0
> +	/* The following test has been disabled due to differences
> +	 * between the upstream and Android kernel drivers. The behaviour
> +	 * that should occur when destroying a timeline with active fences
> +	 * has not been defined yet.
> +	 */
>  	err += RUN_TEST(test_fence_wait_on_destroyed_timeline);
> +#endif

It would be useful to have a real define here that can be
enabled later easily instead of if 0.

thanks,
-- Shuah

>  	err += RUN_TEST(test_stress_two_threads_shared_timeline);
>  	err += RUN_TEST(test_consumer_stress_multi_producer_single_consumer);
>  	err += RUN_TEST(test_merge_stress_random_merge);
> 


-- 
Shuah Khan
Sr. Linux Kernel Developer
Open Source Innovation Group
Samsung Research America (Silicon Valley)
shuahkh@xxxxxxxxxxxxxxx | (970) 217-8978