[PATCH i-g-t v1] tests/kms_async_flips: Optimise crc subtest on simulation platforms

[Karthik B S]

Hi Santhosh,

In the subject line its mentioned crc subtest in specific, but we're 
updating the DPMS subtest as well. Could you please update the subject 
line to capture this, may be just keep it generic.

On 8/26/2025 8:25 AM, Santhosh Reddy Guddati wrote:
> Restrict the crc and dpms subtests to execute on a single pipe on
> the simulation platforms as the crc calculations are slow.
> This avoids timeouts on CI during execution while ensuring
> functionality of the test is verified with single pipe.
>
> Signed-off-by: Santhosh Reddy Guddati <santhosh.reddy.guddati at intel.com>
> ---
>   tests/kms_async_flips.c | 9 +++++++++
>   1 file changed, 9 insertions(+)
>
> diff --git a/tests/kms_async_flips.c b/tests/kms_async_flips.c
> index 733abf417..817de92b9 100644
> --- a/tests/kms_async_flips.c
> +++ b/tests/kms_async_flips.c
> @@ -888,6 +888,13 @@ static void require_linear_modifier(data_t *data)
>   static void run_test(data_t *data, void (*test)(data_t *))
>   {
>   	igt_display_t *display = &data->display;
> +	bool single_pipe = false;
> +	const char *subtest_name = igt_subtest_name();
> +
> +	/* Restrict to one pipe for crc/dpms subtests on simulation platform */
> +	if (igt_run_in_simulation() &&
> +	    (strstr(subtest_name, "crc") || strstr(subtest_name, "async-flip-dpms")))
> +		single_pipe = true;
Although this is right logically, this looks a bit messy. Can we set the 
'single_pipe' may be somewhere during the subtest initialization.
>   
>   	if (data->atomic_path)
>   		require_atomic_async_cap(data);
> @@ -916,6 +923,8 @@ static void run_test(data_t *data, void (*test)(data_t *))
>   			test_init_fbs(data);
>   			test(data);
>   		}
> +		if (single_pipe)
> +			break;

Please ensure to reset the variable, so that we're not affecting other 
subtests unintentionally.

Regards,
Karthik.B.S
>   	}
>   }
>