Re: [igt-dev] [PATCH i-g-t] tests/feature_discovery: detect mst configuration

["Modem, Bhanuprakash" <bhanuprakash.modem@intel.com>]

Hi Swati,

On Sat-26-08-2023 10:35 am, Swati Sharma wrote:
> Add test to detect if we have mst configuration.
> 
> Signed-off-by: Swati Sharma <swati2.sharma@intel.com>
> ---
>   tests/feature_discovery.c | 14 ++++++++++++++

It is out of scope of this patch, but why don't we rename this test to 
kms_feature_discovery? since it is having display specific tests only.

>   1 file changed, 14 insertions(+)
> 
> diff --git a/tests/feature_discovery.c b/tests/feature_discovery.c
> index d1f4b1633..c160a51de 100644
> --- a/tests/feature_discovery.c
> +++ b/tests/feature_discovery.c
> @@ -113,5 +113,19 @@ igt_main {
>   		igt_subtest("psr2") {
>   			igt_require(psr_sink_support(fd, debugfs_fd, PSR_MODE_2));
>   		}
> +
> +		igt_describe("Make sure that we have DP-MST configuration.");
> +		igt_subtest("dp-mst") {
> +			igt_output_t *output;
> +			enum pipe pipe;
> +			int ret;
> +			for_each_valid_output_on_pipe(&display, pipe, output) {
> +				igt_output_set_pipe(output, pipe);

Do we really need to set the output to pipe?
I think, we just need to iterate through connected outputs 
(for_each_connected_output) & check for the encoder (by parsing the 
'PATH' prop blob) is enough.

> +				ret = igt_check_output_is_dp_mst(output);

For me it looks like this change is trying to validate the IGT lib.

I can feel, it would be good to read the 'PATH' property blob & search 
for the encoder name to make sure it is a MST.

- Bhanu

> +				if (ret != 0)
> +					break;
> +			}
> +			igt_require(ret != 0);
> +		}
>   	}
>   }